pandas Notebook Reproducibility and Hidden State Review
Check that a notebook gives the same answer from a clean start.
The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.
[VERIFY] is an editorial flag, not a blank to fill. Wherever the prompt shows [VERIFY], check that item against a reliable source before you use it.
Review a data notebook for reproducibility. Notebook outline with cell order, data sources and paths, random operations, environment and libraries, and outputs shown: [NOTEBOOK_OUTLINE]. Check for cells that depend on execution order, variables defined in deleted cells, absolute paths to personal folders, unseeded randomness, data changed in place, hard-coded dates or filters, missing environment documentation, outputs not matching code, and secrets in cells. Clean-run test: restart the kernel and run all cells in order, then compare key outputs. Worked example: the first run gives a total of 482,300.55 and the clean run 482,301.55, which points to a cell modified after execution. Count issues and rate severity. Return a table with Issue, Evidence, Risk and Fix, and a reproducibility checklist with Check, How, Pass and Result: restart and run all; no absolute paths; seeds set; library versions recorded in an environment file with an example format; input data version recorded; outputs not edited by hand; secrets absent; results match the previous run within tolerance. Propose a structure (parameters at top, functions defined before use, one data load section, setup, load, clean, analyse, report, saved intermediate outputs with dates) and a one-page README outline. Remove outputs with personal data before sharing. Results must be repeatable by someone else; tag version-dependent behaviour [VERIFY].
Fill in before use
Optional: type your details and the prompt above updates. Nothing is saved or sent.
Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.