MI under-age-1 reproducer for #135, with post-fix status (populace#520)#137
Merged
Conversation
Demonstrates the state-specific MI.h5 dataset undercounts Michigan population by ~59% relative to the national dataset filtered by state_code. See #135 for the systemic state-dataset issue. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
…d to populace#520 Re-executed 2026-07-23 with policyengine-us 1.782.0: states/MI.h5 (replaced in the us-data 1.115.4-patch promotion, 2026-05-20) now totals 10,119,131, matching Census. The national default (pinned populace_us_2024 release) puts MI under-1 at 40,204 vs ~104K MDHHS births, with CA at 21.7% of US infants - filed as populace#520. As-filed outputs preserved in #135/#136. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This was referenced Jul 24, 2026
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Adopts #136 by @DTrim99 onto an org branch; supersedes #136. His reproducer commit is cherry-picked with authorship preserved; the ~8.4k lines of unrelated
us/searchlight_institute/files that #136's branch dragged in (they duplicate open PR #126) are excluded — this PR is just the 191-line notebook plus a status addendum.What happened when I ran it (takeover review receipt)
The notebook no longer reproduces the #135 undercount — because the defect was fixed upstream one day after David filed it:
states/MI.h5path)states/MI.h5was replaced on HF in the policyengine-us-data 1.115.4-patch promotion (2026-05-20) and now matches Census on totals. State-specific .h5 datasets undercount population by 30-65% #135's headline defect is resolved.populace_us_2024…20260619), and its under-1 geography is badly off (MI 40.2K vs ~104K births; CA carries 21.7% of US infants vs 11.6% of population; identical shares in the 2024 base year). Filed with full receipts as populace#520.Keeping the notebook merged is deliberate: it's now a standing cross-check of state-vs-national infant counts, and the fastest reproducer for populace#520.
🤖 Generated with Claude Code