Analysis report

Windsor House Prices

Data from R package AER (CRAN), published under https://creativecommons.org/licenses/by/4.0/.

Target
Price
Analyzed
2026-10-28 12:51 UTC

Summary

The analysis modeled Price, using three main inputs out of eleven possible. The model explains 49% of the variation in Price, relying most on Lotsize: moving Lotsize from 3,600 to 6,000 raises the model's output by about 11,520 on average, with everything else as observed. However, the strongest baseline, a straight line through the data, also explains 49%, so the model’s improvement is -0.2 points of variance explained, which is negligible; row by row the two are about level.

Because the model does not beat the straight-line baseline, its output should not be relied on for decisions about individual prices. Caveats include weaker accuracy where Bathrooms, Lotsize, or Prefer are in their highest groups, and flagged differences in errors between groups of Aircon, Driveway, Fullbase, and Prefer. The next step is to validate the model on new data or revisit the modeling approach.

Formula

Price = -3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × BathroomsPrice=−3000+4.8×Lotsize+10000×Stories−30000×−ln2×Bathrooms
Price = -3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathrooms
4 terms · additively decomposable

Held-out diagnostics

TargetQualityErrorR²MAEGeneralizationRows
Price39.34%50.64%0.489914099.9generalizing109
  • Price · Quality 95% CI 15.66% – 52.45%
  • Price · Error 95% CI 39.7% – 70.41%
  • Price · R² 95% CI 0.303 – 0.5828
  • Price · MAE 95% CI 12272.7 – 16280.7

Concerns

WARNING model

The model does not beat a straight line through the data on `Price`

The model explains 49% of the variation on the held-out rows, where a straight line through the data explains 49%: 0.2 points less. The interval of the margin includes zero: on this many held-out rows the two are not told apart. The formula adds little over the reference predictor; whether it is worth its complexity depends on the use. A plain regression on the same inputs does as well or better, so the inputs do carry signal; the search did not find a better formula in the time it had. Row by row the two are about level.

INFO model

Residuals grow in proportion to the predicted value

Across 109 held-out rows split into thirds by predicted value, the residual standard deviation in the top third is 21,931, against 13,423 in the bottom third, 1.63 times as large. The bottom third holds predictions up to 54,354 and the top third predictions from 72,754. Relative to the predictions, the typical miss is 28% of the value in the bottom third and 24% in the top third: the errors grow with the value, as they do for a quantity whose errors scale with its size, and the model is about as accurate at both ends relative to it. Where the same absolute accuracy matters across the range, a log transform of the target evens the spread.

INFO implications

Weaker quality on 3 slices of the data

On 3 of the 24 slices compared the model is weaker than on all held-out rows (R² 0.49). The largest gaps: Bathrooms in the highest quarter (2 to 3) (a typical miss 1.37 times that on all rows, on 28 rows) and Lotsize in the highest quarter (6,360 to 11,440) (a typical miss 1.29 times that on all rows, on 31 rows). Predictions for these rows deserve caution; the Quality by Input Range chapter lists every slice.

Complete list (3)
SegmentRowsR²Against all rowsVerdict
Bathrooms in the highest quarter (2 to 3)280.296-0.194Weaker
Lotsize in the highest quarter (6,360 to 11,440)310.33-0.16Weaker
Prefer = yes250.325-0.165Weaker

Analysis

The report opens with the Headline and the Concerns, followed by the Discovered Relations when the analysis has targets and by Questions & Answers when questions were asked. Three sections carry the detail: Model presents one model per target, Implications states where the models hold and what to weigh before relying on them, and Data describes the table as supplied, column by column, for reference.

Generated 2026-10-28 12:51 UTC, analysis fceb7360-48b7-11f1-8087-d378640fe787, 1 target, time budget 1 h 0 min, compute budget 500 credits, 546 rows. The Provenance & Audit chapter holds the full record.

Validity

Validity: valid. Nothing in the data or the formula speaks against relying on this result.

How to read this

Whether this result is one to rely on. Quality says how closely the formula follows the data; validity says whether that closeness means anything. A formula that only restates its target, or that reads a column which already holds the answer, can reach a high quality and still say nothing new about the data.

What it found

Price starts from a base of -3000, rises with Lotsize, and also changes with Bathrooms and Stories.

  1. Lotsize and Bathrooms carry the model. Lotsize accounts for 51% of what the formula explains and Bathrooms for 32%, with Stories adding 17%.
  2. Price rises with Lotsize. For each additional 1000 units of Lotsize, the model's output for Price is higher by 4800. This holds with the other inputs at typical values.
  3. Aircon is not in the formula. It has the third-strongest association with Price in the table (0.45).
  4. The formula is level with a straight line. The formula explains 49% of Price; a straight line explains 49%, so treat them as level. That was measured on 109 rows it never saw. The formula typically misses Price by 14,100.

Summary

What to do next: Is not told apart from the best baseline, R² lift -0.00175.

6 chapters, each headed by what it found.

S.1 Lotsize carries most of the formula

What the formula relies on. Shares of the formula: Lotsize 51%, Bathrooms 32%, Stories 17%.

The model's output rises with Lotsize on every one of the 14,080 steps between observed values. The model's output rises with Bathrooms on every one of the 660 steps between observed values. The model's output rises with Stories on every one of the 660 steps between observed values. Moving Lotsize from 3,600 to 6,000, its first and third quartile, with every other input as observed raises the model's output by 11,520 on average. These are measurements on the rows the search trained on, inside the observed ranges; outside them nothing is claimed.

Table 4.1.3
InputShare of importancePermutation importanceSensitivityResidual correlationOccurrencesModel's output as the input risesChange, first to third quartile
Lotsize51.29%0.70384.8-0.13631Rises (14,080 of 14,080 steps)+11,520 (3600 to 6000)
Bathrooms31.51%0.432420794.40.21061Rises (660 of 660 steps)
Stories17.2%0.2360100000.063171Rises (660 of 660 steps)+10,000 (1 to 2)
Rows the search trained on, measured
220
Inputs the formula does not read
Aircon, Bedrooms, Driveway, Fullbase, Garage, Gasheat, Prefer, Recreation

Read as a sum, the formula has 4 additive terms: 3 that read an input and a constant base. All that read an input are needed to reach 100% of the variation of the model's output on the 220 rows the search trained on: the largest are -30000 × -ln(2) × Bathrooms (39%), 4.8 × Lotsize (33%) and 10000 × Stories (28%). A share describes the model's output, not the target. Each curve below shows one term's response to its free inputs, with the other inputs held at the training median.

Additive
Yes
Rows the search trained on, measured
220
Terms in the formula
4
Notation
-3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathrooms
Target
Price

S.2 It holds on new rows: R² 0.49 held out, 0.53 on the training rows

How well it holds. The formula explains 49% of Price; a straight line explains 49%, so treat them as level.

How this was checked

  • Measured on 109 rows the formula never saw (20% of 546); the other 437 built it.
  • Compared on the same rows against always guessing the average, always guessing the median and a straight line.
  • Intervals from 1,000 bootstrap resamples of those rows.

R² of 0.49: weak. The model captures only a small part of the variance; either the relevant inputs are missing, or the target depends on interactions the formula did not recover. Mean absolute error is 14,100 on the target's scale.

Error
50.64% (95% CI 39.7% – 70.41%)
MAE
14099.9 (95% CI 12272.7 – 16280.7)
N
109
Training rows
437
Quality
39.34% (95% CI 15.66% – 52.45%)
R²
0.4899 (95% CI 0.303 – 0.5828)
Target
Price

The model does not beat the simple yardsticks: a straight line through the data is ahead by 0.2 points of variance explained on the held-out rows. The model's complexity is not buying anything yet. A plain regression on the same inputs does as well or better, so the inputs do carry signal; the search did not find a better formula in the time it had. Of the 109 held-out rows where the two differ, the model is closer on 50% (95% interval 40% to 59%). The interval of the margin includes zero: on this many held-out rows the two are not told apart.

Best yardstick's score
0.4917 (95% CI 0.3137 – 0.5931)
Best baseline R²
0.4917
Effect size d
0.02714 (95% CI -0.1625 – 0.2151)
Effect size d magnitude
negligible
Margin over the best yardstick
-0.00175 (95% CI -0.06203 – 0.06361)
Margin's interval stays above zero
No
Model MAE
14099.9 (95% CI 12272.7 – 16280.7)
Model
0.4899 (95% CI 0.303 – 0.5828)
Model R²
0.4899 (95% CI 0.303 – 0.5828)
Held-out rows
109
Held-out rows without an answer
0
Target
Price
Rows where the model is closer
49.54% (95% CI 40.33% – 58.78%)
Rows compared
109
Rows level
0
Regression yardstick trained with gaps filled (too few complete rows)
No
Training rows of the regression yardstick
437

The model explains held-out rows about as well as training rows (held-out R² 0.49, gap between training and held-out R² 0.0351): it generalizes. The structure the formula encodes carries from the training rows to the held-out rows.

Generalization gap
0.03513
N test
109
Target
Price
Test error
14099.9 (95% CI 12272.7 – 16280.7)
Test R²
0.4899 (95% CI 0.303 – 0.5828)
Train error
0.4733
Train R²
0.5251
Verdict
Generalizing

S.3 Weaker on 3 slices of the data, most in the highest quarter of Bathrooms

Where it is weak. The largest miss on a held-out row is 68,600 (recorded 175,000, predicted 106,000). 1.8% of the held-out rows are unlike the rows it learned from.

Segment bathrooms in the highest quarter (2 to 3) is mildly weaker than the aggregate (28 held-out rows): segment R² 0.296 against aggregate R² 0.49 (gap -0.194). A noticeable but not severe quality drift; worth watching if this slice represents a high-stakes subset of the production data.

Table 6.3.1
SegmentRowsR²Gap to all rowsVerdict
Bathrooms = q4280.2964-0.1936Weaker than on all rows
Lotsize = q4310.3297-0.1603Weaker than on all rows
Prefer = true250.3245-0.1654Weaker than on all rows
Garage = q4430.4031-0.0869In line with all rows
Fullbase = true390.3460-0.1440In line with all rows
Lotsize = q3260.2179-0.2721In line with all rows
Bedrooms = q4830.4466-0.0434In line with all rows
Stories = q2410.4218-0.0681In line with all rows
Driveway = true900.4420-0.0480In line with all rows
Recreation = false890.5283+0.0384In line with all rows

The first 10 rows of 24. The other 14 are in the notebook.

Aggregate MAE
14099.9
Aggregate R²
0.49
Measure
r2
Target
Price

The worst held-out row was missed by 68,603 (target's scale). The 10 rows with the largest absolute residuals out of 109 held-out rows are shown. Each card breaks the failure down into per-term contributions and input values, so the failure mode is inspectable. A large residual with one concentrated term contribution means one term got this row wrong; a large residual with balanced contributions means a structural mismatch.

Table 4.1.13
ActualPredictedResidualBathroomsLotsizeStories
175,000106,397+68,603299602
155,000105,383+49,617375001
92,50046,802+45,698139601
98,00056,594+41,406160001
34,00071,966-37,966227872
132,00096,029+35,971278002
90,00058,514+31,486164001
75,000105,629-30,629298002
101,00070,418+30,582188801
35,00064,754-29,754177001
Additive
Yes
N extrapolated
0
Held-out rows
109
Held-out rows for the interval
109
Residual P5
-27093
Residual P95
31124
Residual std
18090.6

2 of 109 held-out rows (1.8%) lie further from the training rows than 99 in 100 rows like them would, measured on the 3 inputs the model reads (bathrooms, lotsize, stories). That is a small minority; the held-out figures describe the model on rows like the ones it was trained on.

Table 6.4
TargetInputs measuredLineHeld-out rowsBeyond the lineShare
PriceBathrooms, Lotsize, Stories3.36810921.83%
Table 6.4 · Score bins · 24 of 30 rows shown
CountFromTo
000.135
00.1350.269
00.2690.404
00.4040.539
40.5390.674
30.6740.808
70.8080.943
80.9431.078
181.0781.213
101.2131.347

The first 10 rows of 24. The other 14 are in the notebook.

S.4 Usable with care: 2 points to watch

What to watch. First: The model does not beat a straight line through the data on Price.

Table 2
SeveritySectionConcernDetail
WarningModelThe model does not beat a straight line through the data on PriceThe model explains 49% of the variation on the held-out rows, where a straight line through the data explains 49%: 0.2 points less. The interval of the margin includes zero: on this many held-out rows the two are not told apart. The formula adds little over the reference predictor; whether it is worth its complexity depends on the use. A plain regression on the same inputs does as well or better, so the inputs do carry signal; the search did not find a better formula in the time it had. Row by row the two are about level.
InfoModelResiduals grow in proportion to the predicted valueAcross 109 held-out rows split into thirds by predicted value, the residual standard deviation in the top third is 21,931, against 13,423 in the bottom third, 1.63 times as large. The bottom third holds predictions up to 54,354 and the top third predictions from 72,754. Relative to the predictions, the typical miss is 28% of the value in the bottom third and 24% in the top third: the errors grow with the value, as they do for a quantity whose errors scale with its size, and the model is about as accurate at both ends relative to it. Where the same absolute accuracy matters across the range, a log transform of the target evens the spread.
InfoImplicationsWeaker quality on 3 slices of the dataOn 3 of the 24 slices compared the model is weaker than on all held-out rows (R² 0.49). The largest gaps: Bathrooms in the highest quarter (2 to 3) (a typical miss 1.37 times that on all rows, on 28 rows) and Lotsize in the highest quarter (6,360 to 11,440) (a typical miss 1.29 times that on all rows, on 31 rows). Predictions for these rows deserve caution; the Quality by Input Range chapter lists every slice.

The analysis ran with the configuration below. The checks and the verdict of this chapter are read against it.

  • Targets: 1 target: price
  • Rows: 546 rows
  • Per-Segment Predictions: not produced, because no segment column was chosen
  • Fairness: compared across 6 categorical columns (aircon, driveway, fullbase, gasheat, prefer, recreation); which of them are protected attributes is not decided here
  • Out-of-Distribution: held-out rows are scored against the training rows; predictions made outside this report are not scored
  • Provenance & Audit: recorded in the chapter of that name

Across 6 checks (4 pass, 2 warn, 0 fail), the model passes the hard gates but at least one check warns: Usable with care. A limited audience and monitoring of the warning signals come before a broad rollout.

Table 6.7
NameStatusDetail
GeneralizationPassheld-out quality holds up against training quality
Prediction Interval CoveragePasscoverage matches the interval level, largest gap 3.49 percentage points
Baseline Lift for PriceWarnis not told apart from the best baseline, R² lift -0.00175
Model ConstraintsPassevery held-out prediction is finite and within the training range
FairnessWarn6 columns compared, 4 flagged: Aircon (a group is systematically over- or under-predicted), Driveway (a group receives systematically smaller predictions), Fullbase (a group is systematically over- or under-predicted), Prefer (a group is systematically over- or under-predicted)
Out-of-DistributionPass1.8% of held-out rows are out of distribution
N critical concerns
0
N fail
0
N pass
4
N warn
2
Overall verdict
Caution

Column Driveway splits the held-out rows into 2 groups. The prediction-parity ratio is 0.749 (smallest over largest mean prediction), the largest absolute group residual mean is 4,339 (0.24 times the overall residual standard deviation). Parity gap: at least one group receives systematically smaller predictions on average. Whether that is a property of the target itself or a model bias against the group needs checking.

Aggregate residual std
18090.6
N columns
6
N proxies
0
Proxy threshold
0.5
Table 6.5.1 · Columns
Bias max abs residual meanColumnGroupsInterpretationN groupsPrediction parity ratioVerdict
7548.6Aircontrue, falseColumn Aircon splits the held-out rows into 2 groups. The prediction-parity ratio is 0.846 (smallest over largest mean prediction), the largest absolute group residual mean is 7,549 (0.417 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.8462Residual bias
4339.1Drivewaytrue, falseColumn Driveway splits the held-out rows into 2 groups. The prediction-parity ratio is 0.749 (smallest over largest mean prediction), the largest absolute group residual mean is 4,339 (0.24 times the overall residual standard deviation). Parity gap: at least one group receives systematically smaller predictions on average. Whether that is a property of the target itself or a model bias against the group needs checking.20.7488Prediction parity gap
9036.7Fullbasetrue, falseColumn Fullbase splits the held-out rows into 2 groups. The prediction-parity ratio is 0.989 (smallest over largest mean prediction), the largest absolute group residual mean is 9,037 (0.5 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.9888Residual bias
3795.2Gasheatfalse, trueColumn Gasheat splits the held-out rows into 2 groups. The prediction-parity ratio is 0.91 (smallest over largest mean prediction), the largest absolute group residual mean is 3,795 (0.21 times the overall residual standard deviation). Balanced: every group is within the 80% parity band and no group's residual mean exceeds 30% of the overall residual standard deviation. No systematic disparity was detected on this column.20.9097Balanced
14386.1Prefertrue, falseColumn Prefer splits the held-out rows into 2 groups. The prediction-parity ratio is 0.891 (smallest over largest mean prediction), the largest absolute group residual mean is 14,386 (0.795 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.8914Residual bias
2678.7Recreationtrue, falseColumn Recreation splits the held-out rows into 2 groups. The prediction-parity ratio is 0.967 (smallest over largest mean prediction), the largest absolute group residual mean is 2,679 (0.148 times the overall residual standard deviation). Balanced: every group is within the 80% parity band and no group's residual mean exceeds 30% of the overall residual standard deviation. No systematic disparity was detected on this column.20.9671Balanced

S.5 More rows in the highest quarter of Bathrooms would help most

What to collect next. Its R² there is 0.30 against 0.49 on all rows, on 28 held-out rows.

  • Collect more rows where Bathrooms are in the highest quarter (2 to 3), Lotsize is in the highest quarter (6,360 to 11,440), and Prefer = yes. Reason: the model is less accurate in these segments, explaining 30% of the variation for Bathrooms in the highest quarter, 33% for Lotsize in the highest quarter, and 32% for Prefer = yes.
  • Add data that ensures parity and measures error consistently for Aircon, Driveway, Fullbase, and Prefer groups. Reason: the model's errors differ between groups defined by Aircon, Driveway, Fullbase, and Prefer, as flagged by fairness diagnostics.
  • Collect more data that may improve overall model quality on Price. Reason: the model does not beat a straight line through the data, with only 49% of the variation explained compared with the same 49% for a straight line through the data, and the margin between the two includes zero.

S.6 It shows how Price moves with its inputs in these rows, not what causes it

What this cannot tell you. It was built on Lotsize from 1650 to 16,200, Bathrooms from 1 to 4 and Stories from 1 to 4; outside those ranges it has not been checked. 2 slices of the data had too few held-out rows to judge.

Diafunc found a formula that captures the structure in the training data. That structure is correlative: it states what predicts what in this data. It does not establish that any input causes the target.

A correlative model is the right tool for predicting the target on new rows like these, for exploring which inputs carry information about the target, and for an early-stage diagnostic or screening tool.

A correlative model is not the right tool for claiming that intervening on an input will change the target, for attributing a target's values to specific causes for accountability, or for choosing between policies that act on the inputs.

The Quality by Input Range, Fairness and Concerns chapters show the most common confounding patterns in the data.

The model is correlative: it captures which columns predict the target in the training data, not whether changing one would change the other. Predictions are reliable for inputs that look like the training distribution; they are not evidence that intervening on an input will change the target. The 4 upgrade pathways below show what evidence would support a causal claim.

Table 6.6 · Upgrade pathways
NameSummary
Controlled experimentRandomize the input to be acted on, then analyze again. Randomization breaks confounding, so any change in the target is attributable to the input.
Natural experimentFind a setting where the input varied for reasons unrelated to the target, an instrumental variable. Variation of that kind supports causal inference.
Causal graphWrite out what plausibly causes what, then check whether the model's terms are consistent with the conditional independences the graph implies.
Control for confoundersIf a third variable plausibly moves both the input and the target, stratify by it or condition on it before training again. Gaps in the Quality by Input Range chapter are a clue.

Evidence

Every chapter of the analysis in full: the figures the story rests on, and everything else the analysis measured.

1 Headline

Diafunc modeled Price from 11 columns, of which the model uses 3. The model explains less than half of Price (49%) and relies most on Lotsize, with which the model's output rises.

  • One target was modeled; 109 rows were held out for checking.
  • A straight line through the data already explained 49%; the model did not beat this baseline, with a margin of -0.2 points of variance explained.
  • The model is less accurate on the 31 held-out rows where Lotsize is in the highest quarter (6,360 to 11,440).
Best R²
0.49
Best quality
39.34%
Best target
price
Targets
1

2 Concerns

Table 2
SeveritySectionConcernDetail
WarningModelThe model does not beat a straight line through the data on PriceThe model explains 49% of the variation on the held-out rows, where a straight line through the data explains 49%: 0.2 points less. The interval of the margin includes zero: on this many held-out rows the two are not told apart. The formula adds little over the reference predictor; whether it is worth its complexity depends on the use. A plain regression on the same inputs does as well or better, so the inputs do carry signal; the search did not find a better formula in the time it had. Row by row the two are about level.
InfoModelResiduals grow in proportion to the predicted valueAcross 109 held-out rows split into thirds by predicted value, the residual standard deviation in the top third is 21,931, against 13,423 in the bottom third, 1.63 times as large. The bottom third holds predictions up to 54,354 and the top third predictions from 72,754. Relative to the predictions, the typical miss is 28% of the value in the bottom third and 24% in the top third: the errors grow with the value, as they do for a quantity whose errors scale with its size, and the model is about as accurate at both ends relative to it. Where the same absolute accuracy matters across the range, a log transform of the target evens the spread.
InfoImplicationsWeaker quality on 3 slices of the dataOn 3 of the 24 slices compared the model is weaker than on all held-out rows (R² 0.49). The largest gaps: Bathrooms in the highest quarter (2 to 3) (a typical miss 1.37 times that on all rows, on 28 rows) and Lotsize in the highest quarter (6,360 to 11,440) (a typical miss 1.29 times that on all rows, on 31 rows). Predictions for these rows deserve caution; the Quality by Input Range chapter lists every slice.

How to read this

Data-quality, model-quality, and deployment-risk issues detected during the analysis appear below, ordered by severity. Critical concerns are listed first; Warning and Info entries follow.

3 Discovered Relations

Table 3 · Quality and error measured on the 109 held-out rows
TargetFormulaQualityComplexityError
PricePrice = -3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathrooms39.34%13.2%50.64%

How to read this

One model per target, each a formula with its quality, complexity and error. Higher quality at lower complexity describes the data most clearly. The findings below read the models together; the per-target chapters in the Model section carry each formula, its predictions and its diagnostics.

4 Model

The model explains 49% of the variation in price on held-out rows and includes terms for lot size, stories and bathrooms. Two issues are flagged: the model does not outperform linear regression for price, and residuals increase with predicted value, see Concerns.

Nothing in the Model section's numbers stood out as unusual.

How to read this

The Model section presents one model per target: its formula, its predictions on held-out rows, the inputs it relies on, the baselines it is measured against, and diagnostics of how it behaves. With more than one target, the Composite Model, Multi-Output and Leaderboard chapters read the models as a set.

4.1 Price

The Model section builds one model for Price: a formula that expresses it through the other columns. Diafunc searches the space of mathematical expressions for the formula that balances accuracy against simplicity, so the result is not a black box but an expression that can be read, evaluated and questioned. The chapters that follow show the formula, score it on rows it never saw during training, and take it apart term by term.

4.1.1 Price Model

The model's output relies most on Lotsize, Bathrooms, and Stories: it rises with each. This model described Price with a measured training quality of 26%.

Price = -3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathroomsf:(Bathrooms,Lotsize,Stories)↦PriceBathrooms,Lotsize,Stories,Price∈RPrice=−3000+4.8×Lotsize+10000×Stories−30000×−ln2×Bathrooms
Price = -3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathrooms
Quality 39.34% · Error 50.64% on 109 held-out rows · Complexity 13.2%
Table 4.1.1 · Constants of the formula
ValueRoleInputUnit
-3000InterceptUnit of the model's output
4.8SlopeLotsizeOutput units per unit of Lotsize
10,000SlopeStoriesOutput units per unit of Stories
-30,000Coefficient
Share of the output's variation they carry
100%
Rows the search trained on, measured
220
Rows the search trained on
220
Held out R²
0.4899 (95% CI 0.303 – 0.5828)
Constant base
-3000
Constants in the formula
4
Terms that reach 95% of the output's variation
3
Terms that read an input, constants multiplied in
3
Training rows
437
Held-out rows
109
Inputs the formula reads
3
Terms in the formula
4
Inputs the formula does not read
Aircon, Bedrooms, Driveway, Fullbase, Garage, Gasheat, Prefer, Recreation

How to read this

The three gauges score the model on the held-out rows: rows that were set aside before the search and that the model never saw. They are the figures to quote for how the model does on new data.

Table 4.1.1
MetricMeaning
QualityHow well the model predicts on the held-out rows, marked down for the size of its formula: 100% is an exact prediction, 0% is no better than always predicting the average. It is not R², and sits at or just below it
ErrorHow far the predictions lie from the held-out values, as a share of the miss of always predicting the average: 0% is an exact prediction, 100% is no better than the average
ComplexityThe size of the formula as a share of the complexity the search was allowed: a low figure is a short formula

The figures under During the search are the same measures on the training rows. They are what the search ranked its candidates by, so they describe the search and tend to flatter the model it chose. The Generalization chapter sets the training rows beside the held-out rows, and the Baselines chapter sets the model beside its baselines. The time and credit gauges account for the computation that produced the model, not for the model itself.

4.1.2 Prediction

The model predicts Price for any row that supplies the other columns. This chapter applies it to the held-out rows, the rows set aside before training, so that the predictions can be compared with values the model never saw. The rows themselves and the function that predicts them are kept as reference under Foundation, with the processed data.

4.1.2.1 Prediction Chart

R² of 0.49: weak. The model captures only a small part of the variance; either the relevant inputs are missing, or the target depends on interactions the formula did not recover. Mean absolute error is 14,100 on the target's scale.

Error
50.64% (95% CI 39.7% – 70.41%)
MAE
14099.9 (95% CI 12272.7 – 16280.7)
N
109
Training rows
437
Quality
39.34% (95% CI 15.66% – 52.45%)
R²
0.4899 (95% CI 0.303 – 0.5828)
Target
Price

How to read this

Actual Price on the horizontal axis against the model's predicted value on the vertical axis, for the held-out rows the model never saw during training. Points on the diagonal are perfect predictions; systematic vertical offsets reveal bias, vertical spread reveals noise, and points far from the diagonal are rows the model handles poorly. With 10,000 held-out rows or more, the rows are counted instead of drawn one point each: each cell is a range of the actual and a range of the predicted value, its color the number of rows that fall in both, and the cells whose two ranges match hold the rows the model predicts well.

4.1.2.2 Residual Distribution
Bias
691.7 (95% CI -2389.6 – 4182.1)
N
109
P50
-438.4 (95% CI -4194.4 – 5322.4)
P90
18718.1 (95% CI 14856.8 – 30762.4)
P95
31124 (95% CI 18834.2 – 45697.6)
P99
49303.2 (95% CI 31485.6 – 68603.2)
Std
18090.6 (95% CI 15251.2 – 21116.1)

How to read this

How the held-out residuals (actual Price minus predicted) are distributed. Bias is their mean: a value away from zero means the model is systematically high or low rather than merely imprecise. The standard deviation and the percentiles describe the spread, and the tail percentiles say how large the worst misses get. Every figure carries a bootstrap 95% confidence interval; with a small held-out set those intervals are wide, and that width is itself the finding.

4.1.3 Which inputs matter most

The model's output rises with Lotsize on every one of the 14,080 steps between observed values. The model's output rises with Bathrooms on every one of the 660 steps between observed values. The model's output rises with Stories on every one of the 660 steps between observed values. Moving Lotsize from 3,600 to 6,000, its first and third quartile, with every other input as observed raises the model's output by 11,520 on average. These are measurements on the rows the search trained on, inside the observed ranges; outside them nothing is claimed.

Table 4.1.3
InputShare of importancePermutation importanceSensitivityResidual correlationOccurrencesModel's output as the input risesChange, first to third quartile
Lotsize51.29%0.70384.8-0.13631Rises (14,080 of 14,080 steps)+11,520 (3600 to 6000)
Bathrooms31.51%0.432420794.40.21061Rises (660 of 660 steps)
Stories17.2%0.2360100000.063171Rises (660 of 660 steps)+10,000 (1 to 2)
Rows the search trained on, measured
220
Inputs the formula does not read
Aircon, Bedrooms, Driveway, Fullbase, Garage, Gasheat, Prefer, Recreation

How to read this

The chart ranks the inputs of the Price model by permutation importance: how much the model's error rises when that input's values are shuffled across rows so that it carries no information. A longer bar means the model relies on the input more. The order says which columns the model reads most, and those are the columns where measurement quality and data collection matter most. A short bar marks an input the model barely uses, which is a fact about this model and not a verdict on the column. Importance describes the model's reliance and nothing more: an input can rank high because it stands in for something the data does not contain. Where the search measured it, an input also carries the direction in which the model's output moves as the input rises, counted on the rows the search trained on, and the change of the output when the input alone goes from its first to its third quartile; both describe the formula inside the observed range and say nothing about causes.

4.1.4 Baselines

The model does not beat the simple yardsticks: a straight line through the data is ahead by 0.2 points of variance explained on the held-out rows. The model's complexity is not buying anything yet. A plain regression on the same inputs does as well or better, so the inputs do carry signal; the search did not find a better formula in the time it had. Of the 109 held-out rows where the two differ, the model is closer on 50% (95% interval 40% to 59%). The interval of the margin includes zero: on this many held-out rows the two are not told apart.

Best yardstick's score
0.4917 (95% CI 0.3137 – 0.5931)
Best baseline R²
0.4917
Effect size d
0.02714 (95% CI -0.1625 – 0.2151)
Effect size d magnitude
negligible
Margin over the best yardstick
-0.00175 (95% CI -0.06203 – 0.06361)
Margin's interval stays above zero
No
Model MAE
14099.9 (95% CI 12272.7 – 16280.7)
Model
0.4899 (95% CI 0.303 – 0.5828)
Model R²
0.4899 (95% CI 0.303 – 0.5828)
Held-out rows
109
Held-out rows without an answer
0
Target
Price
Rows where the model is closer
49.54% (95% CI 40.33% – 58.78%)
Rows compared
109
Rows level
0
Regression yardstick trained with gaps filled (too few complete rows)
No
Training rows of the regression yardstick
437

How to read this

The table scores the model for Price against simple yardsticks on the same held-out rows. A yardstick is a predictor simple enough to need no explanation: a constant (the average, the middle value, the most common class or level, the average count), a plain regression on the numeric inputs (a straight line, a logistic regression, a Poisson regression), and for a series over time a forecast from its own past (the last value, the value one season back, the recent trend extended). Which of them appear depends on what kind of column Price is. They are yardsticks only: the analysis has one model, and it is the last row. A regression yardstick answers every row the model answers: where one of its numeric inputs is empty, it reads the average of that input over the training rows. Nothing is filled in for the model. A yardstick whose row is empty could not be scored, and the summary then says what the margin is measured against.

Read the columns from top to bottom. Each kind of target is scored by the measure it is judged on: R² and MAE for a continuous value, with MASE added for a series (below 1 means a typical miss is smaller than the series' usual step from one row to the next); accuracy and balanced accuracy for a yes/no value, with the Brier score where a probability is predicted; accuracy and macro F1 for a class; weighted kappa and the distance in levels for an ordered level; Poisson deviance and RMSE for a count. When the model clears every yardstick by a wide margin, its structure earns its place. When it is level with the regression yardstick, a plain regression describes the data as well. When it is level with the constant, the inputs carry no usable signal for Price.

The margin over the best yardstick comes with an interval. When that interval includes zero, the lead is not settled on this many held-out rows. The share of rows on which the model is closer than the best yardstick is counted only on the rows where the two differ; for a yes/no value or a class these are the rows where the two disagree.

4.1.5 Generalization

The model explains held-out rows about as well as training rows (held-out R² 0.49, gap between training and held-out R² 0.0351): it generalizes. The structure the formula encodes carries from the training rows to the held-out rows.

Generalization gap
0.03513
N test
109
Target
Price
Test error
14099.9 (95% CI 12272.7 – 16280.7)
Test R²
0.4899 (95% CI 0.303 – 0.5828)
Train error
0.4733
Train R²
0.5251
Verdict
Generalizing

How to read this

The chart sets the score of the Price model on the rows it trained on beside its score on the held-out rows it never saw, and the gap between the two bars is the finding. The score is the one that suits the target: R² for a quantity, balanced accuracy for a yes/no target, weighted kappa for ordered levels. The held-out figure carries a bootstrap 95% confidence interval, which says how far it could move with a different held-out draw; the training bar has none, because it is measured on the rows that shaped the model.

The verdict reads from the two figures. Generalizing: the held-out score is strong, or training and held-out agree closely; the structure the formula encodes carries to new rows. Overfit: training is far ahead of held-out and the held-out score is not strong on its own; the formula has absorbed noise from the training rows, and the training figure overstates what to expect on new data. Underfit or noisy: held-out is far ahead of training; this is not strength, it usually means an outlier or a flat region in the training rows that the held-out rows happen to avoid. Insufficient: fewer than 30 held-out rows and no strong held-out score, too little evidence for a negative verdict. A model that predicts the same value for every held-out row gets no verdict: there is nothing that could carry over. In every case the width of the held-out interval says how much weight the gap can bear.

4.1.6 How each term responds to its inputs

Read as a sum, the formula has 4 additive terms: 3 that read an input and a constant base. All that read an input are needed to reach 100% of the variation of the model's output on the 220 rows the search trained on: the largest are -30000 × -ln(2) × Bathrooms (39%), 4.8 × Lotsize (33%) and 10000 × Stories (28%). A share describes the model's output, not the target. Each curve below shows one term's response to its free inputs, with the other inputs held at the training median.

Additive
Yes
Rows the search trained on, measured
220
Terms in the formula
4
Notation
-3000 + 4.8 × Lotsize + 10000 × Stories + 20790 × Bathrooms
Target
Price

How to read this

The formula for Price is split into its additive terms, and each term is traced as one of its inputs moves across the training range while the other inputs stay at their medians. Each curve is therefore a profile of the term at the typical row, not an average over the data: it shows what the term contributes at every value of that input when everything else is ordinary.

A term that depends on several inputs gets one curve per input; a constant term gets none. Terms the formula subtracts are drawn with their sign, so a curve below zero pulls the prediction down. At any fixed row, the terms add up to the model's prediction. Where the search measured it, a term also carries its share of the variation of the model's output on the rows the search trained on; that is a share of what the formula returns, not of Price itself. The notebook lets the reader trim each range to the central 95% of the input. When the formula has no additive structure, the chapter traces the whole formula as a single term along each of its inputs, and the Sensitivity chapter is the better read for how the inputs combine.

4.1.7 How the formula behaves outside the data

The model diverges beyond the observed range of lotsize, stories, bathrooms: predictions there are not supported by the formula's shape. Extrapolate with caution.

Bounded at infinity
No
Rows the search trained on, measured
220
Inputs the formula reads
3
N singularities
0
Risky inputs
lotsize, stories, bathrooms
Target
Price
Table 4.1.7 · Limits at neg infinity
InputKind
LotsizeNegative infinity
StoriesNegative infinity
BathroomsDiverges
Table 4.1.7 · Limits at pos infinity
InputKind
LotsizePositive infinity
StoriesPositive infinity
BathroomsDiverges

How to read this

The formula for Price is read as a mathematical object, without data, to state how it behaves beyond the range it was trained on. A black box cannot answer that question; a formula can. For each input the chapter reports the limit as the input grows without bound in either direction (whether the formula diverges, settles to a finite value or oscillates, with the sign tracked exactly, so that a term like minus two times the cube of the input is reported as diverging downward on the right and upward on the left), whether the formula stays finite in every direction, and any singularity with its solved location: a division whose denominator reaches zero, a logarithm or square root whose argument leaves its domain, or a periodic pole.

The inputs along which the formula diverges or meets a singularity are listed as the risky ones; along them the formula's shape gives no support for extrapolation. A formula that stays finite everywhere with no singularity carries no structural hazard when extrapolated; the data still ends where it ends, and a prediction outside the training range remains unsupported by evidence. The plain-language summary at the end states the same facts in one paragraph.

4.1.8 Contributions per term
Additive
Yes
N rows
5

The notebook draws this chapter's figure from each held-out row; the row-by-row detail is there.

How to read this

For up to five representative held-out rows, the chapter draws one bar chart showing how each additive term of the formula builds that row's prediction of Price. The five rows are the worst prediction (largest absolute residual), the best (smallest), the median miss, the row with the smallest prediction and the row with the largest. When two roles fall on the same row, which is common with few held-out rows, the earlier role keeps it and fewer than five charts appear.

In each chart the bars stack the terms' contributions at the row's input values, terms the formula subtracts sit below zero, and the bars sum to the prediction. The dashed line marks the actual value; the distance between it and the top of the stack is the residual. The bars are evaluated at the same precision as the predictions, so they agree with the Prediction Chart. When the formula has no additive structure, each chart holds one bar for the whole formula, and the Sensitivity chapter is the better read for how individual inputs move it.

4.1.9 What each term is worth

Full-formula R² is 0.49. Each bar below shows how much R² drops when that term is replaced by its mean contribution. Large drop: load-bearing term; near zero: decoration the formula could lose; negative: overfit signal (ablation improves R²).

Additive
Yes
Baseline R²
0.4899
Held-out rows
109

The notebook draws this chapter's figure from each held-out row; the row-by-row detail is there.

How to read this

For each additive term of the formula for Price, the chart shows how much the held-out R² drops when that term is replaced by its mean contribution, so that it no longer varies from row to row. The bars are sorted from the largest drop down, and the dashed line marks the R² of the full model.

A large drop means the term carries real predictive work; the complexity it adds is paid for. A small drop means the term adds little beyond its average value: it may refine a stronger term, or its curve in the chapter "How each term responds to its inputs" may be nearly flat. A negative drop means the held-out R² rises without the term, which is a sign that the term tracks noise in the training rows. Each figure describes what a term contributes inside the model as discovered; the chart diagnoses the model, it is not a guide to removing terms. When the formula has no additive structure, the chart shows one bar for the whole formula.

4.1.10 What-If
Table 4.1.10 · The prediction at the typical row: every input at its median or most common value
ItemValue
Bathrooms1
Lotsize4510
Stories2
Term 1 adds-3000
Term 2 adds21,648
Term 3 adds20,000
Term 4 adds20794.4
Prediction59442.4
Additive
Yes
Held-out rows for the interval
109
Terms in the formula
4
Residual P5
-27093
Residual P95
31124
Residual std
18090.6

How to read this

The chapter shows the model's prediction for Price at one input combination, with the contribution of each term stacked beneath it. The inputs start at their training medians, so the first view is the prediction for a typical row, and each input can move within the range the training rows cover; the notebook lets the reader change the inputs and watch the prediction and the stack update.

In the stack, a long bar marks the term that dominates at this input combination, often the one whose removal costs most in the chapter "What each term is worth". A short bar marks a term that makes a small adjustment here; the chapter "How each term responds to its inputs" shows whether it stays small everywhere or grows toward the edges of the range. A negative bar marks a term that pulls the prediction down at this point, either because the formula subtracts it or because it evaluates below zero.

The inputs stay inside the training range for a reason: the formula computes a value anywhere, but outside that range no training row supports it. When the formula has no additive structure, the stack holds a single bar for the whole formula and the inputs move the prediction directly.

4.1.11 Sensitivity

The prediction responds to 3 inputs. The formula is additive: each curve below shows the change in the prediction per unit of one input (∂ŷ/∂xᵢ) independently. Flat curves mean a constant linear response, bent curves a non-linear response that depends on the position in the range, and zero crossings a response that reverses sign between regions.

Additive
Yes
Terms in the formula
4

How to read this

For each input, the chart traces the rate at which the predicted Price changes as that input moves across its training range, with the other inputs held at their medians. The rate is the local slope of the prediction with respect to the input, measured by a small finite difference at each point. A table summarizes each input with the mean, minimum and maximum of the absolute slope, which ranks the inputs by how strongly the prediction responds to them.

A flat curve means the model is close to linear in that input: the prediction responds at the same rate everywhere in the observed range, which is easy to state and says nothing about values beyond that range. A curve that grows or shrinks marks diminishing or accelerating returns, and whether that is plausible is a question for the domain. A curve that crosses zero means the prediction rises with the input in one part of the range and falls in another, which is either a real optimum or a sign that the formula bends around a few rows. A curve with a large magnitude marks an input to which the prediction is highly responsive; the chapter "What each term is worth" says whether the term behind it also carries predictive work.

In the table, the mean absolute slope ranks the inputs by typical responsiveness. A minimum close to the maximum means the responsiveness is nearly constant; a wide gap means it varies strongly across the range. The chapter reads the whole formula, so it applies whether or not the formula splits into terms.

4.1.12 How two inputs act together

Lotsize and stories enter linearly. No term reads two inputs, so the grid below renders empty; each input's contribution can be read independently in the chapter "How each term responds to its inputs".

Additive
Yes
Inputs in different terms act separately
Yes
Rows the search trained on, measured
220
Inputs the formula is linear in
Lotsize, Stories
Terms in the formula
4
Input pairs not tested
0

How to read this

When a single term of the formula combines two or more inputs, the prediction's response to one of them depends on the value of the other. This chapter draws one surface for each such pair: the predicted Price across the plane of the two inputs, with every other input held at its training median. Pairs that never share a term have no interaction by construction, and the curves in the chapter "How each term responds to its inputs" already say everything about them. Where the search tested a pair, the summary says on how many pairs of rows the two inputs acted together in the model's output; that is a statement about the formula, not a finding that the data hold no other interaction.

Parallel stripes mean the two inputs act almost independently, so the surface is close to a sum of two one-input curves. Curved contours mean a real interaction, where the slope along one input changes with the other; the location of the peak or trough matters. A saddle, rising along one axis and falling along the other, marks a trade-off between the two. A sharp edge or cliff means the formula encodes a change of regime, and the curve for the same input in the chapter "How each term responds to its inputs" shows whether the bend appears there too.

The grid is deliberately coarse, and the color scale runs from low to high predictions. When the formula has no additive structure, the single whole-formula term holds every input, so every pair appears; the number of surfaces is capped for models with many inputs.

4.1.13 Worst Predictions

The worst held-out row was missed by 68,603 (target's scale). The 10 rows with the largest absolute residuals out of 109 held-out rows are shown. Each card breaks the failure down into per-term contributions and input values, so the failure mode is inspectable. A large residual with one concentrated term contribution means one term got this row wrong; a large residual with balanced contributions means a structural mismatch.

Table 4.1.13
ActualPredictedResidualBathroomsLotsizeStories
175,000106,397+68,603299602
155,000105,383+49,617375001
92,50046,802+45,698139601
98,00056,594+41,406160001
34,00071,966-37,966227872
132,00096,029+35,971278002
90,00058,514+31,486164001
75,000105,629-30,629298002
101,00070,418+30,582188801
35,00064,754-29,754177001
Additive
Yes
N extrapolated
0
Held-out rows
109
Held-out rows for the interval
109
Residual P5
-27093
Residual P95
31124
Residual std
18090.6

How to read this

The held-out rows on which the model misses Price by the most, sorted by absolute residual with the largest first and capped at ten. Each card shows the row's actual value, the prediction, the residual, the contribution of every additive term at that row and the row's input values. R² and MAE say how good the model is on average; these cards say where it fails and let each failure be read input by input and term by term.

A large residual with one dominant term means that term produced the wrong value at this row; the row's value for that input, read against its curve in the chapter "How each term responds to its inputs", shows whether it fell in a region the formula models poorly. A large residual with balanced contributions means every term was a little off and the sum still missed, which points to a shape the formula cannot represent in that region. Residuals that change sign from card to card mean the misses are not a consistent bias but scattered errors on particular rows. When the formula has no additive structure, each card shows a single whole-formula contribution.

4.1.14 Prediction Interval Coverage

Across 109 held-out rows, the 95% prediction interval actually covers 94.5% of the rows, within 3.49 percentage points of the promise across every level reported. The per-row prediction interval can be trusted at face value.

Table 4.1.14
Nominal levelHalf-widthEmpirical coverage
50%12201.950.46%
80%23184.183.49%
90%29756.491.74%
95%3545794.5%
Largest miss (percentage points)
3.486
Held-out rows
109
Residual P2.5
-29162.5
Residual P97.5
42693.2
Verdict
well-calibrated

How to read this

Each held-out prediction of Price carries a prediction interval, and this chapter checks whether that interval covers the actual value as often as it claims. The interval reported with each prediction is built from the held-out residuals themselves: its bounds are the 2.5th and 97.5th percentiles of the residuals added to the prediction, so it is asymmetric when the residuals are skewed and assumes no particular error distribution.

The coverage table checks a second, simpler band: the symmetric interval that a normal error distribution with the residuals' standard deviation would give. At four nominal levels (50%, 80%, 90% and 95%) it lists the half-width of that normal band and the share of held-out rows whose actual value falls inside it. Empirical coverage close to nominal means the residuals are close enough to normal that the two bands agree and the interval can be taken at face value. Empirical coverage below nominal means the normal band is too narrow, the residuals have heavier tails than a normal distribution, and the percentile-based interval is the one to trust. Empirical coverage above nominal means the normal band is wider than the data needs, usually because a few large residuals inflate the standard deviation. Every figure comes from the same held-out residuals the intervals are built from.

4.1.15 Regression Diagnostics

109 held-out residuals, with a spread up to 68,603 in absolute terms. Mean residual is 692, close to zero. The model is approximately unbiased on these held-out rows (a small held-out set makes this less reliable; the Prediction Chart chapter states R² and the mean absolute error with their 95% confidence intervals). The charts below check whether the errors look like noise: a pattern in residuals-vs-fitted or a curving Q-Q means a structural miss. The held-out rows are a random draw, so their order carries no time and the autocorrelation of the residuals is not assessed.

Abs residual vs predicted r
0.2836
Bias
691.7
Bias over residual sd
0.03824
Held-out rows
109
Order
shuffled
Residual sd
18090.6
Share within normal band
0.9541

The notebook draws this chapter's figure from each held-out row; the row-by-row detail is there.

How to read this

Four residual diagnostics for Price, all computed from the held-out residuals. Each asks in its own way whether the model's errors look like noise, which is what remains when a model has captured all the structure the data holds.

Residuals against predicted values: Each held-out row is plotted with the model's prediction on the horizontal axis and its residual on the vertical axis, around a zero line. A symmetric cloud around zero means errors of similar size across the predicted range and no systematic bias. A funnel or fan means the size of the errors depends on the predicted value: the model is more uncertain in some regions than in others, and its prediction interval is not equally trustworthy everywhere. A curved pattern, a smile or a frown, means the model misses a structural component; the chapters "How each term responds to its inputs" and "How two inputs act together" show the shapes the formula does capture.

Normal Q-Q plot: The sorted residuals are plotted against the quantiles of a normal distribution, with the diagonal marking a perfectly normal set of residuals. Points that hug the line mean approximately normal residuals, so the normal band in the Prediction Interval Coverage chapter will hold. An S-shape means heavier or lighter tails than normal; heavier tails (below the line at the left, above it at the right) mean the model occasionally produces large errors. A banana or arc means skewed residuals: the model is systematically high or low in part of the range.

Histogram: The residuals binned into equal-width intervals. A bell centered on zero means symmetric errors with no bias. An off-center bell means the model is biased, with a mean error away from zero. A long tail or two peaks mean outliers or two regimes inside the prediction range.

Autocorrelation: For each lag, the correlation of the residual sequence with itself shifted by that many rows, with the band that pure noise would stay inside at plus or minus 1.96 divided by the square root of the row count. Bars inside the band mean uncorrelated residuals with no order-dependent structure left. A bar at lag 1 outside the band means the residuals of neighboring dates correlate: structure in time that the model has not captured. A wave across the lags means periodic structure remains, and the Time Series chapter in the Data section is where to look for its period.

Autocorrelation carries meaning only when the row order itself does. When the data has a date column, the held-out rows are the most recent ones in date order, and the autocorrelation is assessed. Without a date column the held-out rows are a random draw: the row after a held-out row is an arbitrary other row, so the autocorrelation is not assessed and the summary above says so.

4.1.16 Search Progress

The search ran for 71 generations, ending at quality 0.376 (gain +0.32 from generation 2). The curve shows a substantial climb. The last improvement came after 39% of the search time: the search had settled well before it stopped, so more time alone is unlikely to change the result.

First generation
2
First quality
5.62%
Last generation
72
Last improvement ms
1,389,349
Last quality
37.59%
Max quality
37.59%
N points
71
N total generations
71
Search time ms
3,600,224
Stop reason
Maximum runtime reached

How to read this

The line traces the best quality the search for the Price model had reached at each generation, from the first generation to the last. Quality is the technology's combined score on the training rows, the figure the Model Metrics chapter defines. A curve that flattens well before the last generation means the search converged: the budget was more than enough, and the model is the one the search settled on. A curve still rising at the last generation means the budget ended before the search settled, and the model is the best the search reached within it. The chart describes the search, not the data: it says whether the search had time to settle, not whether more rows would help.

4.1.17 Model Constraints

Across 109 held-out predictions, the formula is mathematically well-behaved: every output is finite, stays inside the training-target ±2σ band, and no prediction runs away beyond the bounded-prediction limit. It is safe to consume downstream at face value.

Table 4.1.17
NameStatusDetail
Output coherentPassno non-finite predictions across 109 rows
Within domainPassno rows outside the training target's ±2σ range across 109 rows
Bounded predictionsPassno runaway predictions beyond ±10σ across 109 rows
Held-out rows
109
N non finite
0
N outside domain
0
N runaway
0
Verdict
Clean

How to read this

Sanity checks on the Price formula, run against the held-out rows. Three checks: every prediction is a finite number (output coherent), predictions stay within the range of the training target widened by two standard deviations on each side (within domain), and no single prediction runs away more than ten standard deviations from the training mean (bounded predictions).

These are pass or fail guards, not quality measures: a formula can clear all three and still predict poorly. A failure is the stronger signal. It means the formula misbehaves numerically somewhere in the data it was shown, and any downstream use of it should wait until that is understood.

4.1.18 Modeling

Measured during the search, on the rows the model was trained on

MAE of a straight line through the inputs (same rows)
12520.4
RMSE of a straight line through the inputs (same rows)
17167.5
Training MAE (rows the search trained on)
12789.7
MAE of always predicting the mean (same rows)
21560.5
RMSE of always predicting the mean (same rows)
28416.9
Training R² (rows the search trained on)
0.5665
Training RMSE (rows the search trained on)
18710.4
Rows the search trained on, measured
220
Error
47.33%
Inputs the straight line used
5
N train
437
Quality
25.52%

How to read this

The function below trains the model for Price on the processed data. Its output is the model that every other chapter of this section scores and dissects.

4.2 Findings

2 findings from the model diagnostics (1 Warning, 1 Info) are listed in the Concerns chapter at the top of the analysis.

5 Data

Several columns correlate with price, and levels of aircon and prefer each account for a measurable share of the spread. Patterns related to lot size and bathrooms are present as correlations and group averages.

Most notable in the Data section:

  • Lotsize and Price move together (correlation 0.536)
  • Aircon accounts for 20% of the spread in Price: yes averages 25,996 above no (95% interval 21,212 to 30,780)
  • Bathrooms and Price move together (correlation 0.517)

How to read this

The Data section describes the table as supplied, before any modeling: the Column Profiles overview of every column and its role, one chapter per column with its distribution and correlations, a closer look at columns with notable characteristics, Duplicate Rows, Leakage Suspicion, a map for paired coordinate columns, and the General chapter with descriptive statistics, correlations, missing values, outlier ranges, clusters and segments across all columns.

5.1 Column Profiles

Profiled 12 columns across 546 rows (6 numeric, 6 non-numeric): every column is at least 75% populated. Distributional statistics on numeric columns can be read at face value. 4 columns are strongly skewed: the mean of such a column is pulled toward its long tail, so the median is the better guide to a typical value.

Table 5.1 · Columns
NameRoleTypeNMissingMeanStdMinP25MedianP75MaxSkewnessCV
PriceTargetinteger546068,12226,70325,00049,12562,00082,000190,0001.210.392
AirconInputbool5460
BathroomsInputinteger54601.2860.502111241.5920.3906
BedroomsInputinteger54602.9650.737123360.49590.2487
DrivewayInputbool5460
FullbaseInputbool5460
GarageInputinteger54600.69230.8613000130.84481.244
GasheatInputbool5460
LotsizeInputinteger546051502168165036004600636016,2001.3230.421
PreferInputbool5460

The first 10 rows of 12. The other 2 are in the notebook.

N columns
12
N high missing
0
N rows
546
N strongly skewed
4
Table branch
main
Table ID
3513c229-baee-4d47-a43b-65907217eae8

How to read this

One row per column of the analyzed table: its role in the analysis (the target being predicted, an input it is predicted from, or the segment column), its type, value count and missing count, and for numeric columns the mean, standard deviation, minimum, quartiles, maximum and interquartile range. A column with 25% or more missing values is flagged; its mean and quartiles describe only the rows that have a value.

Two further figures describe the shape of a numeric column. The skewness says whether the values lean to one side: close to zero the column is symmetric, from 0.5 it is moderately skewed and from 1 strongly skewed, with the longer tail on the high side when the figure is positive and on the low side when it is negative. In a strongly skewed column the mean is pulled toward the tail, and the median is the better guide to a typical value. The coefficient of variation is the standard deviation divided by the mean: it compares how widely columns spread regardless of their units, and is given only for a column whose mean is positive.

These are the columns as supplied, before any modeling: the reference against which every later chapter reads.

5.2 Aircon

Aircon: 2 distinct values, most often no.

How to read this

A column chapter profiles one column: its distribution in figures and charts, and how it relates to the other columns. It opens with one line on what the column holds. The Description chapter carries its figures, with every other column summarized for each of its values where it has few, and its association with every other column where it is numeric. The Visualization chapter carries its charts. A table with a date column adds an Over Time and a Time Series chapter; a target with classes adds a Class Balance chapter.

5.2.1 Description
Table 5.2.1 · Aircon
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modefalse
Table 5.2.1 · Rows per value
VariableValueRowsShare
Airconfalse37368.32%
Aircontrue17331.68%

How to read this

The figures below summarize Aircon in numbers: where its values center (mean, median, mode), how far they spread (range, variance, standard deviation, interquartile range) and how they are distributed across percentiles. They are the reference for every chart in the Visualization chapter and for the correlation measures further down.

The figures for Aircon are grouped: the column itself first, then where its values center, how far they spread, the shape of their distribution and its percentiles. Each statistic is explained in the Description chapter under General in the Data section.

5.2.1.1 By Value

Aircon accounts for 20% of the spread in Price: yes averages 25,996 above no (95% interval 21,212 to 30,780). An association, not a cause.

Table 5.2.1.1 · Groups
TargetValueRowsAverage or share95% interval
Pricetrue17385880.681609.1 to 90152.1
Pricefalse37359884.957712.3 to 62057.4
Table 5.2.1.1 · Mean per value of Aircon
AirconRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false3732.8854823.11.63359884.90.60051.223
true1733.1395855.62.18585880.60.89021.422

How to read this

The rows are grouped by the value of Aircon, and every other column is summarized within each group with the same descriptive statistics. A statistic that shifts from group to group marks a column that moves with Aircon; one that stays level marks a column that does not. Groups with few rows produce unstable figures, so the row count of each group is the first thing to read. Where the groups differ in a target by more than chance would give, the chapter states by how much; it states that the two go together, not that one brings the other about.

5.2.2 Visualization

The charts in this chapter were chosen from the type of Aircon and of the columns it is compared with: a numeric column gets a histogram, a box plot and, when its row order carries a pattern, a value sequence, a categorical column gets a frequency chart, and its relation to each target of the analysis, and to the segment column when there is one, gets the chart that suits the pair, with the target on the vertical axis where it is a number. A column is not charted against every other column; which columns move together is shown under Strongest Correlations. The charts complement the figures in the Description chapter; the shape of a distribution and the position of its outliers are easier to see than to read.

5.3 Bathrooms

Bathrooms: 1 to 4, median 1, skewed to high values, 4 distinct values; moves with Price most (r 0.517).

5.3.1 Description
Table 5.3.1 · Bathrooms
StatisticValue
Typeinteger
Values present546
Distinct values4
Continuousno
Decimal places0
Minimum1
Maximum4
Mean1.286
Median1
Mode1

The first 10 rows of 17. The other 7 are in the notebook.

Table 5.3.1 · Rows per value
VariableValueRowsShare
Bathrooms140273.63%
Bathrooms213324.36%
Bathrooms3101.832%
Bathrooms410.1832%
5.3.1.1 By Value

Bathrooms accounts for 25% of the spread in Price: 3 averages 43,877 above 1 (95% interval 14,194 to 73,559). An association, not a cause.

Table 5.3.1.1 · Groups
TargetValueRowsAverage or share95% interval
Price310104,03074382.7 to 133,677
Price213388702.984006.9 to 93399
Price140260153.258146.2 to 62160.3
Table 5.3.1.1 · Mean per value of Bathrooms
BathroomsRowsBedroomsLotsizeStoriesPriceGarage
14022.8034921.11.63460153.20.6095
21333.3835699.52.29388702.90.8947
3103.866772.1104,0301.1
41489604175,0003
5.3.1.2 Correlation
Table 5.3.1.2 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Pricetwo numbers0.51670.51670.4524 to 0.57570.47870.392< 0.001546
Bedroomstwo numbers0.37690.37380.2992 to 0.44380.37690.3548< 0.001546
Storiestwo numbers0.32410.32410.2469 to 0.39720.30660.2875< 0.001546
Lotsizetwo numbers0.21360.19380.1117 to 0.27330.21360.1744< 0.001546
Aircona number and a category0.185< 0.001546
Garagetwo numbers0.17820.17820.0957 to 0.25820.17480.1648< 0.001546
Recreationa number and a category0.12690.004676546
Fullbasea number and a category0.10280.02237546
Gasheata number and a category0.067370.1416546
Prefera number and a category0.064010.1487546

The first 10 rows of 11. The other 1 is in the notebook.

How to read this

Each row pairs Bathrooms with one other column, numeric or categorical, and reports how strongly the two go together on an association from 0 to 1, and whether an association this strong could be chance. For another numeric column the row adds the direction: Pearson's coefficient measures a straight-line relationship, and Spearman's rho and Kendall's tau measure a monotonic relationship on ranks, which makes them resistant to outliers and lets them see a curved but consistently rising or falling pattern. For a categorical column the association is the correlation ratio η, which has no direction. A strong association is not an explanation: two columns can go together when a third one moves both.

The table lists every other numeric or categorical column against Bathrooms, with the pair's kind and association, three correlation coefficients and a confidence interval where the other column is numeric, the q-value and the rows the pair was measured on.

Table 5.3.1.2
MeasureMeaning
KindWhat the other column holds: a number (numeric) or categories (numeric_categorical). A categorical column is one that is not numeric and holds between 2 and 50 distinct values
AssociationThe strength of the pair from 0 to 1. With a numeric column: the stronger of Pearson's coefficient and Spearman's ρ, without its sign. With a categorical column: the correlation ratio η, the square root of the share of Bathrooms's spread that lies between the categories
CovarianceNumeric partner only. The direction of the relationship (positive or negative) in the units of the two columns. It is not standardized, so its size says little about strength without the columns' own spread
Pearson's Correlation CoefficientNumeric partner only. The strength and direction of a linear relationship, from -1 (a perfect falling line) through 0 (no linear relationship) to 1 (a perfect rising line)
Spearman's ρNumeric partner only. The strength and direction of a monotonic relationship, computed on ranks rather than values, from -1 to 1. It follows a curved pattern as long as it rises or falls consistently, and outliers move it little
Kendall's τNumeric partner only. The strength and direction of association from the agreement and disagreement of rank order across pairs of rows, from -1 to 1. It is the steadier of the two rank measures when the rows are few
q-valueHow likely a pair this strong is among unrelated columns, adjusted for the number of pairs involving Bathrooms (Benjamini-Hochberg). Among all pairs reported with q below 0.05, about 5% or fewer are expected to be false discoveries
95% CI (Pearson)Numeric partner only. The lower and upper bounds of the 95% confidence interval on the Pearson coefficient, from the Fisher z-transform. An interval that contains 0 leaves the direction of the relationship open
RowsThe rows in which both columns hold a value. The q-value and the interval of a pair rest on this count
5.3.2 Visualization

The rows of Bathrooms are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of 0.03 between each value and the next), so the order carries no pattern to show.

5.4 Bedrooms

Bedrooms: 1 to 6, median 3, 6 distinct values; moves with Stories most (r 0.408).

5.4.1 Description
Table 5.4.1 · Bedrooms
StatisticValue
Typeinteger
Values present546
Distinct values6
Continuousyes
Decimal places0
Minimum1
Maximum6
Mean2.965
Median3
Mode3

The first 10 rows of 17. The other 7 are in the notebook.

5.4.1.1 By Value

Bedrooms accounts for 15% of the spread in Price: 5 averages 31,254 above 2 (95% interval 5,622 to 56,886). An association, not a cause.

Table 5.4.1.1 · Groups
TargetValueRowsAverage or share95% interval
Price51083,14057567.9 to 108,712
Price49581853.775137.8 to 88569.6
Price330170817.267989.3 to 73645.2
Price21365188649453.6 to 54318.4
Table 5.4.1.1 · Mean per value of Bedrooms
BedroomsRowsLotsizeStoriesPriceGarageBathrooms
12371038,750
21364636.21.169518860.49261.059
33015225.91.93770817.20.72091.266
4955582.12.30581853.70.91581.621
5106291.5283,1400.61.8
623950268,4501.5
5.4.1.2 Correlation
Table 5.4.1.2 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Storiestwo numbers0.50330.4080.3355 to 0.47560.50330.4493< 0.001546
Pricetwo numbers0.39030.36640.2915 to 0.43690.39030.3067< 0.001546
Bathroomstwo numbers0.37690.37380.2992 to 0.44380.37690.3548< 0.001546
Lotsizetwo numbers0.17630.15190.06882 to 0.23280.17630.1356< 0.001546
Aircona number and a category0.1604< 0.001546
Garagetwo numbers0.14650.13910.05586 to 0.22050.14650.13080.002049546
Fullbasea number and a category0.09720.03633546
Recreationa number and a category0.080490.07976546
Prefera number and a category0.078950.07976546
Gasheata number and a category0.046030.3113546

The first 10 rows of 11. The other 1 is in the notebook.

5.4.2 Visualization

The rows of Bedrooms are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of 0.06 between each value and the next), so the order carries no pattern to show.

5.5 Driveway

Driveway: 2 distinct values, most often yes.

5.5.1 Description
Table 5.5.1 · Driveway
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modetrue
Table 5.5.1 · Rows per value
VariableValueRowsShare
Drivewaytrue46985.9%
Drivewayfalse7714.1%
5.5.1.1 By Value

Driveway accounts for 8.6% of the spread in Price: yes averages 22,778 above no (95% interval 18,997 to 26,559). An association, not a cause.

Table 5.5.1.1 · Groups
TargetValueRowsAverage or share95% interval
Pricetrue46971333.968881.5 to 73786.3
Pricefalse7748555.845654.6 to 51456.9
Table 5.5.1.1 · Mean per value of Driveway
DrivewayRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false772.9873606.41.54548555.80.25971.234
true4692.9625403.71.85171333.90.76331.294

5.6 Fullbase

Fullbase: 2 distinct values, most often no.

5.6.1 Description
Table 5.6.1 · Fullbase
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modefalse
Table 5.6.1 · Rows per value
VariableValueRowsShare
Fullbasefalse35565.02%
Fullbasetrue19134.98%
5.6.1.1 By Value
Table 5.6.1.1 · Mean per value of Fullbase
FullbaseRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false3552.9135074.81.91864477.60.65921.248
true1913.0635290.51.60274894.50.75391.356

5.7 Garage

Garage: 0 to 3, median 0, skewed to high values, 4 distinct values; moves with Price most (r 0.383).

5.7.1 Description
Table 5.7.1 · Garage
StatisticValue
Typeinteger
Values present546
Distinct values4
Continuousno
Decimal places0
Minimum0
Maximum3
Mean0.6923
Median0
Mode0

The first 10 rows of 17. The other 7 are in the notebook.

Table 5.7.1 · Rows per value
VariableValueRowsShare
Garage030054.95%
Garage112623.08%
Garage210819.78%
Garage3122.198%
5.7.1.1 By Value

Garage accounts for 15% of the spread in Price: 2 averages 25,071 above 0 (95% interval 18,772 to 31,370). An association, not a cause.

Table 5.7.1.1 · Groups
TargetValueRowsAverage or share95% interval
Price210884233.378322.3 to 90144.2
Price31283816.754677.6 to 112,956
Price112674148.469413.2 to 78883.6
Price030059162.356954.4 to 61370.3
Table 5.7.1.1 · Mean per value of Garage
GarageRowsBedroomsLotsizeStoriesPriceBathrooms
03002.8874502.81.76359162.31.207
11262.9845576.41.89774148.41.349
21083.136226.41.80684233.31.407
3123.257178.3283816.71.5
5.7.1.2 Correlation
Table 5.7.1.2 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Pricetwo numbers0.38330.38330.3093 to 0.45270.36360.2815< 0.001546
Lotsizetwo numbers0.35290.35290.2772 to 0.42420.35260.2749< 0.001546
Drivewaya number and a category0.2037< 0.001546
Bathroomstwo numbers0.17820.17820.0957 to 0.25820.17480.1648< 0.001546
Aircona number and a category0.1566< 0.001546
Bedroomstwo numbers0.14650.13910.05586 to 0.22050.14650.13080.002049546
Prefera number and a category0.092360.04861546
Gasheata number and a category0.068140.1536546
Fullbasea number and a category0.052520.2694546
Storiestwo numbers0.043410.04341-0.04065 to 0.12690.016180.014590.3424546

The first 10 rows of 11. The other 1 is in the notebook.

5.7.2 Visualization

The rows of Garage are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of -0.00 between each value and the next), so the order carries no pattern to show.

5.8 Gasheat

Gasheat: 2 distinct values, most often no.

5.8.1 Description
Table 5.8.1 · Gasheat
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modefalse
Table 5.8.1 · Rows per value
VariableValueRowsShare
Gasheatfalse52195.42%
Gasheattrue254.579%
5.8.1.1 By Value
Table 5.8.1.1 · Mean per value of Gasheat
GasheatRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false5212.9585154.61.80467579.10.67951.278
true253.125059.31.8879,4280.961.44

5.9 Lotsize

Lotsize: 1,650 to 16,200, median 4,600, skewed to high values; moves with Price most (r 0.536).

5.9.1 Description
Table 5.9.1 · Lotsize
StatisticValue
Typeinteger
Values present546
Distinct values284
Continuousyes
Decimal places0
Minimum1650
Maximum16,200
Mean5150.3
Median4600
Mode6000

The first 10 rows of 17. The other 7 are in the notebook.

5.9.1.1 Correlation
Table 5.9.1.1 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Pricetwo numbers0.6030.53580.4732 to 0.5930.6030.4225< 0.001546
Garagetwo numbers0.35290.35290.2772 to 0.42420.35260.2749< 0.001546
Drivewaya number and a category0.2888< 0.001546
Prefera number and a category0.2348< 0.001546
Aircona number and a category0.2218< 0.001546
Bathroomstwo numbers0.21360.19380.1117 to 0.27330.21360.1744< 0.001546
Bedroomstwo numbers0.17630.15190.06882 to 0.23280.17630.1356< 0.001546
Recreationa number and a category0.14030.001389546
Storiestwo numbers0.083680.08368-0.000241 to 0.16640.058810.048240.06195546
Fullbasea number and a category0.047490.2948546

The first 10 rows of 11. The other 1 is in the notebook.

5.9.2 Visualization

The rows of Lotsize are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of -0.04 between each value and the next), so the order carries no pattern to show.

5.10 Prefer

Prefer: 2 distinct values, most often no.

5.10.1 Description
Table 5.10.1 · Prefer
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modefalse
Table 5.10.1 · Rows per value
VariableValueRowsShare
Preferfalse41876.56%
Prefertrue12823.44%
5.10.1.1 By Value

Prefer accounts for 11% of the spread in Price: yes averages 20,723 above no (95% interval 15,256 to 26,189). An association, not a cause.

Table 5.10.1.1 · Groups
TargetValueRowsAverage or share95% interval
Pricetrue12883986.479029.1 to 88943.7
Pricefalse41863263.560935.4 to 65591.5
Table 5.10.1.1 · Mean per value of Prefer
PreferRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false4182.9334868.81.78763263.50.64831.268
true1283.076069.31.87583986.40.83591.344

5.11 Price

Price: 25,000 to 190,000, median 62,000, skewed to high values; moves with Lotsize most (r 0.536).

5.11.1 Description
Table 5.11.1 · Price
StatisticValue
Typeinteger
Values present546
Distinct values219
Continuousyes
Decimal places0
Minimum25,000
Maximum190,000
Mean68121.6
Median62,000
Mode50000

The first 10 rows of 17. The other 7 are in the notebook.

5.11.1.1 Correlation
Table 5.11.1.1 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Lotsizetwo numbers0.6030.53580.4732 to 0.5930.6030.4225< 0.001546
Bathroomstwo numbers0.51670.51670.4524 to 0.57570.47870.392< 0.001546
Aircona number and a category0.4533< 0.001546
Storiestwo numbers0.42120.42120.3496 to 0.48790.36470.2877< 0.001546
Bedroomstwo numbers0.39030.36640.2915 to 0.43690.39030.3067< 0.001546
Garagetwo numbers0.38330.38330.3093 to 0.45270.36360.2815< 0.001546
Prefera number and a category0.3291< 0.001546
Drivewaya number and a category0.2972< 0.001546
Recreationa number and a category0.255< 0.001546
Fullbasea number and a category0.1862< 0.001546

The first 10 rows of 11. The other 1 is in the notebook.

5.11.2 Visualization

The rows of Price are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of 0.07 between each value and the next), so the order carries no pattern to show.

5.12 Recreation

Recreation: 2 distinct values, most often no.

5.12.1 Description
Table 5.12.1 · Recreation
StatisticValue
Typebool
Values present546
Distinct values2
Continuousno
Modefalse
Table 5.12.1 · Rows per value
VariableValueRowsShare
Recreationfalse44982.23%
Recreationtrue9717.77%
5.12.1.1 By Value

Recreation accounts for 6.3% of the spread in Price: yes averages 17,796 above no (95% interval 12,277 to 23,314). An association, not a cause.

Table 5.12.1.1 · Groups
TargetValueRowsAverage or share95% interval
Pricetrue9782755.777780.9 to 87730.4
Pricefalse44964960.162540.7 to 67379.5
Table 5.12.1.1 · Mean per value of Recreation
RecreationRowsBedroomsLotsizeStoriesPriceGarageBathrooms
false4492.93850091.79164960.10.67711.256
true973.0935804.31.88782755.70.76291.423

5.13 Stories

Stories: 1 to 4, median 2, skewed to low values, 4 distinct values; moves with Price most (r 0.421).

5.13.1 Description
Table 5.13.1 · Stories
StatisticValue
Typeinteger
Values present546
Distinct values4
Continuousno
Decimal places0
Minimum1
Maximum4
Mean1.808
Median2
Mode2

The first 10 rows of 17. The other 7 are in the notebook.

Table 5.13.1 · Rows per value
VariableValueRowsShare
Stories122741.58%
Stories223843.59%
Stories3407.326%
Stories4417.509%
5.13.1.1 By Value

Stories accounts for 18% of the spread in Price: 4 averages 43,397 above 1 (95% interval 36,184 to 50,610). An association, not a cause.

Table 5.13.1.1 · Groups
TargetValueRowsAverage or share95% interval
Price441102,97896239 to 109,717
Price34081,24073284.1 to 89195.9
Price223868058.264545.2 to 71571.2
Price122759580.856918.9 to 62242.7
Table 5.13.1.1 · Mean per value of Stories
StoriesRowsBedroomsLotsizePriceGarageBathrooms
12272.525289.859580.80.70931.137
22383.2824756.968058.20.65131.332
3403.2255308.781,2400.451.35
4413.3416506.2102,9781.0731.78
5.13.1.2 Correlation
Table 5.13.1.2 · Correlations
ColumnPair ofAssociationPearson's r95% CI (Pearson)Spearman's ρKendall's τq-valueRows
Bedroomstwo numbers0.50330.4080.3355 to 0.47560.50330.4493< 0.001546
Pricetwo numbers0.42120.42120.3496 to 0.48790.36470.2877< 0.001546
Bathroomstwo numbers0.32410.32410.2469 to 0.39720.30660.2875< 0.001546
Aircona number and a category0.2962< 0.001546
Fullbasea number and a category0.1739< 0.001546
Drivewaya number and a category0.12250.007606546
Lotsizetwo numbers0.083680.08368-0.000241 to 0.16640.058810.048240.07965546
Garagetwo numbers0.043410.04341-0.04065 to 0.12690.016180.014590.3565546
Prefera number and a category0.042940.3565546
Recreationa number and a category0.042280.3565546

The first 10 rows of 11. The other 1 is in the notebook.

5.13.2 Visualization

The rows of Stories are not charted in their order: consecutive rows are no more alike than rows picked at random (a correlation of -0.02 between each value and the next), so the order carries no pattern to show.

5.14 Duplicate Rows

Across 546 rows the table has 2 exact-duplicate rows (0.4% of the table; biggest duplicate group 2 rows). Below 1% is typically benign, occasional rounding-encoded collisions or a repeated import, but worth a spot-check before downstream modeling.

Largest group size
2
N exact duplicates
2
N exact groups
1
N rows
546
N unique
545
Pct exact duplicates
0.37
Verdict
Minor

How to read this

This chapter counts rows that repeat another row exactly, value for value: how many rows take part in a duplicate group, how many groups exist, and how large the largest group is.

Duplicate rows inflate the apparent row count and can place identical rows on both sides of the training and held-out split, which flatters every quality figure in this report. Only exact repeats are detected here; rows that differ by a rounding digit or a whitespace variant count as distinct.

5.15 Leakage Suspicion

Nothing to check. No column that awaits an answer tracks a target at 0.95 or more, so no column on its own gives a target away.

Categorical verdict
Clean
Critical threshold
0.999
N categorical critical
0
N categorical suspects
0
N confirmed
0
N critical
0
N suspects
0
Threshold
0.95
Verdict
Clean

How to read this

For every column that is not a target, this chapter reads how closely the column tracks each target, on an association from 0 to 1, and lists any pair at 0.95 or more. For two numeric columns the association is the stronger of Pearson's r and Spearman's rho; when the column or the target is categorical it is the correlation ratio η, or Cramér's V when both are. At 0.999 or more the column reproduces the target: that is what a copy or a rescaled copy of it looks like.

A listed column poses one question: is its value known before the target's value exists? If it is, the column is a strong input and the quality stated in this report is real. If it was computed from the target or recorded after it, a formula that reads it looks right on these rows and will not hold on new data. The coefficients read the same in both cases; the answer needs someone who knows where the column came from.

The owner of the analysis can give that answer. A column confirmed as known before the target is listed as confirmed and is no longer a suspect. A column that turns out to be a leak is left out of the model's inputs, and the analysis run again.

5.16 General

Clustering found no clear groups. The best grouping of the 546 complete rows over 5 numeric columns is into 2 groups, with a silhouette of 0.31 (on a scale where 1 is fully separate groups and 0 is none), below the 0.50 that clear structure shows, so no Clusters chapter is written.

Table 5.16 · Groupings tried
GroupsSilhouetteInertiaKept
20.31052014yes
30.28621640.3
40.27191467.6
50.29561347.6
60.28081199
Inertia
2014
K
2
N columns used
5
N rows
546
N rows complete
546
Silhouette
0.3105
Verdict
Weak

How to read this

The General chapter covers what holds across all columns together: descriptive statistics, correlations, missing values and outlier ranges under Description, the Clusters found among the rows, and, when a segment column was chosen, the Segments.

5.16.1 Description
Table 5.16.1 · Columns
VariableTypeValues presentDistinct valuesContinuousDecimal places
Bedroomsinteger5466yes0
Lotsizeinteger546284yes0
Drivewaybool5462no
Storiesinteger5464no0
Gasheatbool5462no
Preferbool5462no
Priceinteger546219yes0
Recreationbool5462no
Garageinteger5464no0
Airconbool5462no

The first 10 rows of 12. The other 2 are in the notebook.

Table 5.16.1 · Center
VariableMeanMedianMode
Bedrooms2.96533
Lotsize5150.346006000
Drivewaytrue
Stories1.80822
Gasheatfalse
Preferfalse
Price68121.662,00050000
Recreationfalse
Garage0.692300
Airconfalse

The first 10 rows of 12. The other 2 are in the notebook.

Table 5.16.1 · Spread
VariableMinimumMaximumStandard deviationVarianceInterquartile rangeSum
Bedrooms160.73740.543711619
Lotsize165016,2002168.24,700,91227602,812,045
Driveway
Stories140.86820.75381987
Gasheat
Prefer
Price25,000190,00026702.7713,032,63532,87537,194,392
Recreation
Garage030.86130.74181378
Aircon

The first 10 rows of 12. The other 2 are in the notebook.

Table 5.16.1 · Shape
VariableSkewnessKurtosisTotal sum of squares
Bedrooms0.49590.735296.3
Lotsize1.3232.7632.562e+09
Driveway
Stories1.0750.6562410.8
Gasheat
Prefer
Price1.211.963.886e+11
Recreation
Garage0.8448-0.5686404.3
Aircon

The first 10 rows of 12. The other 2 are in the notebook.

Table 5.16.1 · Rows per value
VariableValueRowsShare
Drivewaytrue46985.9%
Drivewayfalse7714.1%
Stories122741.58%
Stories223843.59%
Stories3407.326%
Stories4417.509%
Gasheatfalse52195.42%
Gasheattrue254.579%
Preferfalse41876.56%
Prefertrue12823.44%

The first 10 rows of 24. The other 14 are in the notebook.

How to read this

This chapter summarizes every column at once: the descriptive statistics table below, then Correlation, Missing Values and Outliers. For a numeric column the table reports its center (mean, median, mode), its spread (standard deviation, variance, interquartile range, percentiles) and its shape (skewness, kurtosis); for a categorical column it reports the count and share of each category.

Each row of the tables describes one column. The statistics are grouped: the column itself, where its values center, how far they spread, what shape their distribution takes, and its percentiles. They read as follows.

Table 5.16.1
StatisticExplanation
VariableName of the column the statistics in the row describe
TypeType assigned during data processing (bool, integer, number, decimal, text or date)
Values presentNumber of rows in which the column holds a value
Distinct valuesNumber of different values the column holds
ContinuousWhether the column can take any value in a range, as opposed to a discrete column limited to a fixed set of values
Decimal placesLength of the fractional part of the values
MeanArithmetic average: the sum of all values divided by the number of values
MedianThe middle value: half of the values lie at or below it and half at or above it. It is also the 50% percentile
ModeThe value that appears most often
MinimumSmallest numeric value in the column

The first 10 rows of 20. The other 10 are in the notebook.

5.16.1.1 Missing Values

No column has a missing value: all 546 rows are complete.

N columns with missing
0
N complete rows
546
N rows
546
Pct complete rows
100

How to read this

The table lists, for each column, the count of present values (Values) beside the row count of the processed table (Rows), the number of missing values and their share. The larger the share, the less any statistic or model built on the column can be trusted.

  • Below 5% missing: safe to use as it is.
  • 5% to 25% missing: usable; filling the gaps or excluding those rows is worth considering.
  • 25% or more missing: the column is flagged and receives its own chapter in the Data section. Verify why so much is missing before trusting any result that depends on it.

A model is computed only on rows that hold a value in every column it reads, so gaps spread over several columns cost more rows than any one column's share suggests. The last column, Complete rows without, shows for each input with gaps how many rows would be complete if that column were left out; it is blank for a column without gaps, for a target, and for a column no model reads. A column that is the only gap in many rows is the one whose removal keeps the most data.

5.16.1.2 Outliers

6 of 6 numeric columns have values beyond the Tukey fences (1.5 times the interquartile range outside the quartiles). stories has the largest share: 7.5% of its values. Whether such values are rare but real rows or recording errors decides whether they are kept.

Table 5.16.1.2 · Outliers
ColumnValuesLower fenceUpper fenceBelowAboveShare
Bathrooms546none below3.5000010.18%
Bedrooms546none below4.50000122.2%
Garage546none below2.5000122.2%
Lotsize546none below10,5000122.2%
Price546none below131,3130152.75%
Stories546none below3.50000417.51%
N columns
6
N columns assessed
6
N columns with outliers
6
Worst column
stories
Worst pct
7.51

How to read this

The table counts the outliers of every numeric column by the Tukey rule. Q1 and Q3 are the values below which 25% and 75% of a column's values fall, and the interquartile range is Q3 minus Q1. The lower fence lies 1.5 times the interquartile range below Q1, the upper fence the same distance above Q3; a value beyond a fence is an outlier. For each column the table gives both fences, how many values lie below and above them, their share of the column's values, and the single value that lies furthest beyond its fence. These are the same fences the box plots draw their whiskers to.

A column is not assessed when it has fewer than four values, or when the middle half of its values sits on a single number: the interquartile range is then zero and the rule would call every other value an outlier.

Extreme values can be real (rare but legitimate rows) or errors (measurement mistakes, data-entry slips). Which case applies decides whether to keep, clip or drop them before trusting a model that uses the column. A strongly skewed column has many values beyond one fence without any of them being wrong.

5.16.1.3 Correlation

66 column pairs analyzed across 546 rows (15 of two numbers, 36 of a number and a category, 15 of two categories); 43 (65%) are unlikely to be chance, after allowing for how many pairs were tested. Any other pair may still show a large association, but one that cannot be told apart from chance on the rows the pair was measured on. Every pair is stated as an association from 0 to 1, whatever its kind, so pairs of different kinds can be ranked against each other. An association states that two columns go together, not that one brings the other about.

N categorical pairs
15
N mixed pairs
36
N numeric pairs
15
N pairs
66
N samples
546
N significant
43
Table 5.16.1.3 · Pairs · 24 of 66 rows shown
Association95% CIColumnEta squaredGroupsKendallKindNOther columnPearsonQ valueSpearman
0.6030.4732, 0.593Lotsize0.4225numeric546Price0.5358< 0.0010.603
0.51670.4524, 0.5757Bathrooms0.392numeric546Price0.5167< 0.0010.4787
0.50330.3355, 0.4756Bedrooms0.4493numeric546Stories0.408< 0.0010.5033
0.4533Aircon0.20552numeric_categorical546Price< 0.001
0.42120.3496, 0.4879Price0.2877numeric546Stories0.4212< 0.0010.3647
0.39030.2915, 0.4369Bedrooms0.3067numeric546Price0.3664< 0.0010.3903
0.38330.3093, 0.4527Garage0.2815numeric546Price0.3833< 0.0010.3636
0.37690.2992, 0.4438Bathrooms0.3548numeric546Bedrooms0.3738< 0.0010.3769
0.3703Fullbase2, 2categorical546Recreation< 0.001
0.35290.2772, 0.4242Garage0.2749numeric546Lotsize0.3529< 0.0010.3526

The first 10 rows of 24. The other 14 are in the notebook.

How to read this

This chapter measures how pairs of columns go together: the Correlation Measures table below, the Strongest Correlations, the Correlation Heatmap and, when the analysis has targets, the Potential Data Leaks. Numeric columns and categorical columns are both paired, a categorical column being one that is not numeric and holds between 2 and 50 distinct values. Every pair is stated as an association from 0 (unrelated) to 1 (one column gives the other away), so pairs of different kinds can be compared. An association states that two columns go together, not that one brings the other about.

The table lists every pair of columns once, with its kind, its association, the figures of its kind, its q-value and the number of rows the pair was measured on. A pair is measured on the rows where both columns hold a value.

Table 5.16.1.3
MeasureExplanation
KindWhat the two columns hold: two numbers (numeric), a number and a category (numeric_categorical), or two categories (categorical)
AssociationStrength of the pair from 0 to 1, whatever its kind. Two numbers: the stronger of Pearson's coefficient and Spearman's ρ, without its sign. A number and a category: the correlation ratio η. Two categories: Cramér's V
CovarianceTwo numbers only. Direction of the relationship (positive or negative) in the units of the two columns. It is not standardized, so its size says little about strength without further context
Pearson's Correlation CoefficientTwo numbers only. Strength and direction of a linear relationship, from -1 (perfect negative) through 0 (no linear relationship) to 1 (perfect positive)
Spearman's ρTwo numbers only. Strength and direction of a monotonic relationship, computed on ranks rather than values. It captures relationships that consistently rise or fall even when they are not linear, and it is less affected by extreme values
Kendall's τTwo numbers only. Strength and direction of a monotonic relationship based on the concordance of ranks between the two columns. More conservative than Spearman's ρ, and steadier on small tables
Correlation ratio ηA number and a category. The square root of the share of the number's spread that lies between the categories, from 0 (every category has the same average) to 1 (the category fixes the number). For a category of two values it equals the absolute Pearson coefficient with the category coded 0 and 1. Its square, η², is that share: an η of 0.6 means the categories account for 36% of the spread
Cramér's VTwo categories. How far the table of counts of the two columns departs from what unrelated columns would give, from 0 (unrelated) to 1 (each value of one column goes with one value of the other), corrected so that columns with many rare values do not read as related
q-valueHow likely a pair this strong is among unrelated columns, adjusted for the number of pairs listed (Benjamini-Hochberg false-discovery rate). Among all pairs reported with q below 0.05, about 5% or fewer are expected to be false discoveries. The right measure for ranking many pairs
95% CI (Pearson)Two numbers only. Lower and upper bounds of the 95% confidence interval on the Pearson coefficient, via the Fisher z-transform, on the rows the pair was measured on. A narrow interval means the coefficient is well estimated; a wide interval that straddles zero means the relationship is uncertain

The first 10 rows of 12. The other 2 are in the notebook.

5.16.1.3.1 Strongest Correlations

The 20 pairs of columns with the smallest q-values: the associations that remain significant after the Benjamini-Hochberg adjustment for the number of pairs tested. Among all pairs reported with q below 0.05, about 5% or fewer are expected to be false discoveries; a pair above 0.05 may be coincidence even when its association looks strong.

A pair with an association above 0.9 often signals a duplicate or derived column. Moderately strong pairs with a low q-value are the relationships most worth investigating.

Each listed pair is drawn below, the picture behind the figure. Two numbers are drawn as a scatter, one point per row, the quickest way to see whether a strong value rests on a trend or on a few rows far from the rest. A number and a category are drawn as the number's box plot per category. Two categories are drawn as a heat map of how many rows hold each combination of values.

5.16.1.3.3 Potential Data Leaks

The table lists the association between every column and each target, whatever the two hold, with its q-value. An association close to 1 with a target, together with a q-value near zero, is one of two things: a strong input, or a data leak, a column derived from the target after the fact, such as a tax amount computed from the price being predicted. A leaked column trains a model that looks perfect and fails on new data, because the leak is not available at prediction time.

The Leakage Suspicion chapter in the Data section lists the pairs at 0.95 or more and marks those at 0.999 or more, where a column reproduces the target. For two numbers the association is whichever of Pearson's r and Spearman's rho is the stronger, because a column derived from the target often follows a curve rather than a line: the log of a price tracks the price perfectly, and Pearson alone reads that as weaker than it is. For a number and a category it is the correlation ratio η, and for two categories Cramér's V. Whether a listed column is a strong input or a leak depends on whether it is known before the target exists, which the owner of the analysis can confirm.

6 Implications

Accuracy declines for rows with higher values of bathrooms, lotsize, and for prefer marked yes, so the pattern in price is not captured as well in these cases. One concern is flagged, see the Concerns chapter.

Most notable in the Implications section:

  • The model is less accurate on rows with Bathrooms in the highest quarter (2 to 3) (19 points less variance explained than overall)
  • The model is less accurate on rows with Lotsize in the highest quarter (6,360 to 11,440) (16 points less variance explained than overall)
  • The model is less accurate on rows with Prefer = yes (17 points less variance explained than overall)

How to read this

The Implications section states how to act on the models: the Stakeholder Summary and the Recommended Next Data, where the models hold and where they weaken (Quality by Input Range, Rows unlike the ones the model learned from, Per-Segment Predictions, Fairness), what to weigh before relying on them (Causal Honesty, Deployment Readiness), and the Provenance & Audit trail a reviewer needs to retrace the analysis.

6.1 Stakeholder Summary

The analysis modeled Price, using three main inputs out of eleven possible. The model explains 49% of the variation in Price, relying most on Lotsize: moving Lotsize from 3,600 to 6,000 raises the model's output by about 11,520 on average, with everything else as observed. However, the strongest baseline, a straight line through the data, also explains 49%, so the model’s improvement is -0.2 points of variance explained, which is negligible; row by row the two are about level.

Because the model does not beat the straight-line baseline, its output should not be relied on for decisions about individual prices. Caveats include weaker accuracy where Bathrooms, Lotsize, or Prefer are in their highest groups, and flagged differences in errors between groups of Aircon, Driveway, Fullbase, and Prefer. The next step is to validate the model on new data or revisit the modeling approach.

6.2 Recommended Next Data

  • Collect more rows where Bathrooms are in the highest quarter (2 to 3), Lotsize is in the highest quarter (6,360 to 11,440), and Prefer = yes. Reason: the model is less accurate in these segments, explaining 30% of the variation for Bathrooms in the highest quarter, 33% for Lotsize in the highest quarter, and 32% for Prefer = yes.
  • Add data that ensures parity and measures error consistently for Aircon, Driveway, Fullbase, and Prefer groups. Reason: the model's errors differ between groups defined by Aircon, Driveway, Fullbase, and Prefer, as flagged by fairness diagnostics.
  • Collect more data that may improve overall model quality on Price. Reason: the model does not beat a straight line through the data, with only 49% of the variation explained compared with the same 49% for a straight line through the data, and the margin between the two includes zero.

6.3 Quality by Input Range

Quality by Input Range splits the held-out rows by every usable input and compares the model's quality on each segment with its quality on all held-out rows. For a number the quality is R², the share of variation explained; for a yes-or-no target it is accuracy, the share of rows classified correctly, the measure the Classification Performance chapter reports. A numeric input is split into quartiles; a categorical input with at most ten values is split by value. A segment with fewer than three held-out rows is left out, because its gap would be noise.

The chart is sorted from the worst gap to the best. Each bar is the segment's figure minus the figure on all held-out rows.

  • Negative bars: the model is weaker on these rows than on all held-out rows. A segment 0.25 or more below the overall value (25 points of accuracy) reads much weaker; a segment 0.10 to 0.25 below reads weaker. These are the segments to treat with caution.
  • Positive bars: the model is stronger on these rows than overall. A strong segment can mask weakness elsewhere, so the negative segments come first.
6.3.1 Price by Input Range

Segment bathrooms in the highest quarter (2 to 3) is mildly weaker than the aggregate (28 held-out rows): segment R² 0.296 against aggregate R² 0.49 (gap -0.194). A noticeable but not severe quality drift; worth watching if this slice represents a high-stakes subset of the production data.

Table 6.3.1
SegmentRowsR²Gap to all rowsVerdict
Bathrooms = q4280.2964-0.1936Weaker than on all rows
Lotsize = q4310.3297-0.1603Weaker than on all rows
Prefer = true250.3245-0.1654Weaker than on all rows
Garage = q4430.4031-0.0869In line with all rows
Fullbase = true390.3460-0.1440In line with all rows
Lotsize = q3260.2179-0.2721In line with all rows
Bedrooms = q4830.4466-0.0434In line with all rows
Stories = q2410.4218-0.0681In line with all rows
Driveway = true900.4420-0.0480In line with all rows
Recreation = false890.5283+0.0384In line with all rows

The first 10 rows of 24. The other 14 are in the notebook.

Aggregate MAE
14099.9
Aggregate R²
0.49
Measure
r2
Target
Price

6.4 Rows unlike the ones the model learned from

2 of 109 held-out rows (1.8%) lie further from the training rows than 99 in 100 rows like them would, measured on the 3 inputs the model reads (bathrooms, lotsize, stories). That is a small minority; the held-out figures describe the model on rows like the ones it was trained on.

Table 6.4
TargetInputs measuredLineHeld-out rowsBeyond the lineShare
PriceBathrooms, Lotsize, Stories3.36810921.83%
Table 6.4 · Score bins · 24 of 30 rows shown
CountFromTo
000.135
00.1350.269
00.2690.404
00.4040.539
40.5390.674
30.6740.808
70.8080.943
80.9431.078
181.0781.213
101.2131.347

The first 10 rows of 24. The other 14 are in the notebook.

How to read this

Out-of-distribution detection finds held-out rows that lie far from the training rows of a model. It looks only at the numeric inputs the model reads. Each of them is standardized against the mean and standard deviation of the training rows, and the standardized values are combined into one distance per row. The line a row is read against depends on the number of inputs: it is the distance that 99 in 100 rows like the training rows stay below. A row beyond the line is flagged as out of distribution, and the held-out figures are unlikely to apply to it. When the held-out rows resemble the training rows, about one row in a hundred is flagged.

The chart shows one bar per held-out row in row order; the flagged rows are the bars at or above the line. With more than one target, the chart shows the model with the largest flagged share, and the figures name it.

The check treats inputs independently, so it will not notice a rare combination of values that are each ordinary on their own, such as a large height with a small weight. It is a first screen. A decision that depends on catching unusual inputs needs a dedicated check.

6.5 Fairness

This chapter checks the held-out predictions for uneven treatment of groups. For every categorical column with two to ten distinct values, other than the target, it reports per group the row count, mean actual value, mean prediction and mean residual, with R² and MAE for a number or accuracy, the share of rows classified correctly, for a yes-or-no target, and two measures across the groups.

  • Prediction parity ratio: the smallest group mean prediction divided by the largest, in absolute value. A ratio of 1.0 is perfect parity. Below 0.80, the four-fifths rule, some group receives systematically different predictions.
  • Largest mean residual: the largest absolute mean residual of any group. Above 30% of the overall residual standard deviation, at least one group is systematically over- or under-predicted.

The verdict combines the two: balanced when neither measure trips, parity gap or residual bias when one does, parity gap and residual bias when both do, and insufficient data when fewer than two groups have enough held-out rows.

Diafunc does not decide which columns are protected attributes; it reports every plausible one. A flagged column matters where it is a protected attribute in the decision the model supports, and can be set aside where it is not.

6.5.1 Price Fairness

Column Driveway splits the held-out rows into 2 groups. The prediction-parity ratio is 0.749 (smallest over largest mean prediction), the largest absolute group residual mean is 4,339 (0.24 times the overall residual standard deviation). Parity gap: at least one group receives systematically smaller predictions on average. Whether that is a property of the target itself or a model bias against the group needs checking.

Aggregate residual std
18090.6
N columns
6
N proxies
0
Proxy threshold
0.5
Table 6.5.1 · Columns
Bias max abs residual meanColumnGroupsInterpretationN groupsPrediction parity ratioVerdict
7548.6Aircontrue, falseColumn Aircon splits the held-out rows into 2 groups. The prediction-parity ratio is 0.846 (smallest over largest mean prediction), the largest absolute group residual mean is 7,549 (0.417 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.8462Residual bias
4339.1Drivewaytrue, falseColumn Driveway splits the held-out rows into 2 groups. The prediction-parity ratio is 0.749 (smallest over largest mean prediction), the largest absolute group residual mean is 4,339 (0.24 times the overall residual standard deviation). Parity gap: at least one group receives systematically smaller predictions on average. Whether that is a property of the target itself or a model bias against the group needs checking.20.7488Prediction parity gap
9036.7Fullbasetrue, falseColumn Fullbase splits the held-out rows into 2 groups. The prediction-parity ratio is 0.989 (smallest over largest mean prediction), the largest absolute group residual mean is 9,037 (0.5 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.9888Residual bias
3795.2Gasheatfalse, trueColumn Gasheat splits the held-out rows into 2 groups. The prediction-parity ratio is 0.91 (smallest over largest mean prediction), the largest absolute group residual mean is 3,795 (0.21 times the overall residual standard deviation). Balanced: every group is within the 80% parity band and no group's residual mean exceeds 30% of the overall residual standard deviation. No systematic disparity was detected on this column.20.9097Balanced
14386.1Prefertrue, falseColumn Prefer splits the held-out rows into 2 groups. The prediction-parity ratio is 0.891 (smallest over largest mean prediction), the largest absolute group residual mean is 14,386 (0.795 times the overall residual standard deviation). Residual bias: at least one group is systematically over- or under-predicted. The model's errors are not zero-mean across groups; its predictions for the affected groups deserve caution.20.8914Residual bias
2678.7Recreationtrue, falseColumn Recreation splits the held-out rows into 2 groups. The prediction-parity ratio is 0.967 (smallest over largest mean prediction), the largest absolute group residual mean is 2,679 (0.148 times the overall residual standard deviation). Balanced: every group is within the 80% parity band and no group's residual mean exceeds 30% of the overall residual standard deviation. No systematic disparity was detected on this column.20.9671Balanced

6.6 Causal Honesty

Diafunc found a formula that captures the structure in the training data. That structure is correlative: it states what predicts what in this data. It does not establish that any input causes the target.

A correlative model is the right tool for predicting the target on new rows like these, for exploring which inputs carry information about the target, and for an early-stage diagnostic or screening tool.

A correlative model is not the right tool for claiming that intervening on an input will change the target, for attributing a target's values to specific causes for accountability, or for choosing between policies that act on the inputs.

The Quality by Input Range, Fairness and Concerns chapters show the most common confounding patterns in the data.

The model is correlative: it captures which columns predict the target in the training data, not whether changing one would change the other. Predictions are reliable for inputs that look like the training distribution; they are not evidence that intervening on an input will change the target. The 4 upgrade pathways below show what evidence would support a causal claim.

Table 6.6 · Upgrade pathways
NameSummary
Controlled experimentRandomize the input to be acted on, then analyze again. Randomization breaks confounding, so any change in the target is attributable to the input.
Natural experimentFind a setting where the input varied for reasons unrelated to the target, an instrumental variable. Variation of that kind supports causal inference.
Causal graphWrite out what plausibly causes what, then check whether the model's terms are consistent with the conditional independences the graph implies.
Control for confoundersIf a third variable plausibly moves both the input and the target, stratify by it or condition on it before training again. Gaps in the Quality by Input Range chapter are a clue.

6.7 Deployment Readiness

The analysis ran with the configuration below. The checks and the verdict of this chapter are read against it.

  • Targets: 1 target: price
  • Rows: 546 rows
  • Per-Segment Predictions: not produced, because no segment column was chosen
  • Fairness: compared across 6 categorical columns (aircon, driveway, fullbase, gasheat, prefer, recreation); which of them are protected attributes is not decided here
  • Out-of-Distribution: held-out rows are scored against the training rows; predictions made outside this report are not scored
  • Provenance & Audit: recorded in the chapter of that name

Across 6 checks (4 pass, 2 warn, 0 fail), the model passes the hard gates but at least one check warns: Usable with care. A limited audience and monitoring of the warning signals come before a broad rollout.

Table 6.7
NameStatusDetail
GeneralizationPassheld-out quality holds up against training quality
Prediction Interval CoveragePasscoverage matches the interval level, largest gap 3.49 percentage points
Baseline Lift for PriceWarnis not told apart from the best baseline, R² lift -0.00175
Model ConstraintsPassevery held-out prediction is finite and within the training range
FairnessWarn6 columns compared, 4 flagged: Aircon (a group is systematically over- or under-predicted), Driveway (a group receives systematically smaller predictions), Fullbase (a group is systematically over- or under-predicted), Prefer (a group is systematically over- or under-predicted)
Out-of-DistributionPass1.8% of held-out rows are out of distribution
N critical concerns
0
N fail
0
N pass
4
N warn
2
Overall verdict
Caution

6.8 Provenance & Audit

This record captures the inputs and configuration that produced the report, so the analysis can be reproduced or the result defended in an audit.

Table 6.8
ItemValue
Analysis IDfceb7360-48b7-11f1-8087-d378640fe787
Run time2026-10-28 13:56 UTC
Total rows546
Modeled targetsPrice
Segment columnnot available
Time limit1 h 0 min
Time taken1 h 3 min
Credit budget500 credits
Ufinq version0.3.48

6.9 Findings

1 finding from the implications diagnostics (1 Info) is listed in the Concerns chapter at the top of the analysis.

Loading Diafunc…