Create intermediate Works with: Any AI

Data Quality Report From a DataFrame

Rubric scoring Has example

Produce a repeatable profile of a dataset.

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

Write a data quality report for a pandas DataFrame. Dataset description: [DATASET_DESCRIPTION]. Output of df.info() and df.describe(include="all"): [PASTE_SUMMARIES]. Business rules the data should obey: [BUSINESS_RULES].

Return:
1. A table per column: Column | Type | Non-null % | Distinct values | Min / max or top values | Issues found | Severity (blocking, important, cosmetic).
2. Rule checks written as code, one per business rule, each returning the offending rows (for example negative quantities, ship date before order date, totals that do not equal quantity times price), with counts.
3. Distribution checks: outliers by IQR rule and by domain limits, and how to tell a real outlier from an error.
4. Consistency checks across columns and across tables.
5. A scorecard: completeness, validity, uniqueness, consistency, timeliness, each scored 1 to 5 with the evidence. Show how you computed each score.
6. A prioritized fix list with owners (data producer versus analyst).
Make the checks reusable as a function that returns a dataframe of failures. Say where you are guessing about meaning.

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.

Shape: 12,480 rows x 14 columns.
Column | Type | Missing % | Unique | Issues
customer_id | int | 0.0 | 12,480 | OK, unique
email | object | 4.1 | 11,900 | 580 duplicates, 12 without @
signup_date | object | 0.0 | 640 | stored as text
age | float | 6.3 | 82 | 9 values over 110
Rules checked: age between 0 and 110 (9 failures); email contains @ (12 failures).
Duplicates: 0 exact rows; 580 repeated emails.
Top fixes: convert signup_date; review ages; deduplicate by email with a rule agreed with the owner.
Re-run the report after cleaning and compare.

• data-quality