pandas Missing Data Handling Audit
Check how nulls are treated and what that does to results.
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.
Audit missing data handling in a pandas workflow. Columns with percent missing, the code that fills or drops them, and the analysis that follows: [MISSING_DATA_AND_CODE]. Describe missingness before treating it: a table of each column with count missing, percent missing, whether missing values cluster by group or date, what the data owner says they mean, current treatment, effect and recommendation. Check for blanket fillna with zero, dropna removing too many rows (show count and percentage lost), imputing with the mean before splitting data for modelling, missing values coded as strings or special numbers like 999, and missingness that may carry meaning. Decision guide: if missing is random and small, dropping may be acceptable; if related to the outcome, imputation can bias results; if 0 or 999 means unknown, convert it to a missing marker first. Effect arithmetic: 50 known values averaging 50 and 10 unknown filled with 0 give a mean of 2,500 / 60 = 41.7 instead of 50, a 16.7 percent understatement. Rule: document every treatment in the notebook, add a missingness summary table to the output, and report results with and without the treatment when the difference matters. Provide a template methods paragraph on how missing data were identified, handled and tested. Do not hide missingness. I will run and check everything myself; label what you cannot be sure of [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.