Audit intermediate Works with: Any AI

Missing Data Strategy Review

Decision tree Has example

Decide how to handle blanks responsibly.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

Help me decide how to handle blanks in my dataset responsibly. Dataset and purpose of the analysis: [DATASET_AND_PURPOSE]. Columns with missing values and the share missing in each: [MISSING_SUMMARY]. How the data was collected: [COLLECTION_METHOD]. Tool: [TOOL].

For each affected column, run this decision tree and report the outcome:
1. Why is it missing? Not collected, not applicable, refused, error, or truly zero? Blank is not the same as zero.
2. Is the pattern random, or related to other variables (for example higher-income respondents skip the income question)? Suggest a check.
3. How much is missing? Under 5 percent, 5 to 20 percent, over 20 percent: different treatment.
4. Options: leave blank and exclude from calculations, drop rows, fill with a placeholder category such as "Unknown", fill with the median or mode (with warning), use the previous value (for time series), estimate from other columns, or go back to the source.
5. What does each option do to the results I report? State the effect on averages, totals, and counts.
6. How will I document the choice?

Output a table: Column | Missing % | Likely reason | Chosen treatment | Effect on results | Documentation note. Add formulas for my tool and a caveat paragraph for the report.

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.

Column | Missing % | Likely reason | Treatment | Effect on results | Note
Income | 23% | Respondents skipped (sensitive); related to age group | Keep as "Not stated", report separately | Average income only for those who answered | Do not fill with the median
Delivery date | 3% | Entry error | Go back to source; fix 14 rows | None after fix | Documented
Satisfaction score | 8% | Not applicable (no purchase) | Exclude from the average | Average based on 920 respondents | State n
Caveat for the report: "Income data are missing for 23 percent of respondents and may differ from reported values." Formulas: =COUNTBLANK(C2:C1000)/ROWS(C2:C1000).

• missing-data
• imputation