docs(tutorials): fix broken Hillstrom tutorial code - #137
Merged
Conversation
Two issues that prevented the tutorial from running end-to-end: - Treatment mapping used 'Women E-Mail', but the CSV segment label is 'Womens E-Mail'. As a result D=2 was NaN for the Women's Email arm and every subsequent Women's-vs-anything analysis silently produced no samples. - The overall-population PTE plots used revenue_locations[1:] as the x-axis (50 points) against a PTE vector of length 51 returned by predict_pte(len(locations)-1), causing a broadcast shape mismatch. Dropping the slice matches the (correct) convention already used in the subgroup section. Verified by running the tutorial end-to-end on the live Hillstrom CSV; ATE numbers now match the "Expected output" block in the tutorial. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
1 task
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
While verifying the tutorials referenced in the JOSS paper's Research Impact Statement, two bugs in docs/docs/tutorials/hillstrom.md prevented it from running end-to-end.
Summary
'Women E-Mail'but the CSV segment label is'Womens E-Mail'. As a resultD=2wasNaNfor the Women's Email arm and every subsequent Women's-vs-anything analysis silently produced zero samples for that arm.revenue_locations[1:](50 points) against a PTE vector of length 51 returned bypredict_pte(len(locations)-1), which crashed with a broadcast-shape error insideax.bar. Dropping the slice matches the (correct) convention already used in the subgroup section further down the tutorial.Verified by running the tutorial end-to-end against the live Hillstrom CSV; the ATE numbers now match the "Expected output" block in the tutorial exactly.
Test plan