Map and inspect an unfamiliar data export

Quickly understand the purpose, column types, and data quality risks of a messy raw CSV export. Ideal when inheriting unfamiliar operations sheets.

You getDataset Summary, column Dictionary and immediate Audit Steps

Prompt

Prompt
Act as a senior data analyst. Review the sample rows and header names pasted below to help me understand this unfamiliar dataset.
- Infer the business purpose of the file and define each column plainly.
- Flag probable primary keys, date formats, and numeric types.
- List potential data quality issues, missing values, or edge cases to inspect before doing analysis.
Output format:
1. Dataset Summary (2 sentences on likely source and purpose)
2. Column Dictionary (table with columns: Column Name, Inferred Data Type, Description, Expected Values)
3. Immediate Audit Steps (bulleted checklist of what to check in the raw data)
Here is the data snippet:

Example output

Example for: an e-commerce order export with messy columns

  1. Dataset Summary
    This dataset appears to be an e-commerce transaction log tracking individual customer purchases and fulfillment statuses. It is likely sourced directly from an order management or billing database.
  2. Column Dictionary
Column NameInferred Data TypeDescriptionExpected Values
ord_idString / KeyUnique identifier for each checkout orderNon-empty alphanumeric string
user_refIntegerForeign key representing the registered buyerPositive integer
created_dtDateTimeTimestamp when order was placedISO timestamp or YYYY-MM-DD HH:MM
net_amtFloatFinal billed amount after discountsNumeric >= 0.00
currString (3)Three-letter currency codeUSD, EUR, GBP, CAD
status_codeIntegerNumerical indicator of fulfillment status1 (Paid), 2 (Shipped), 3 (Cancelled)
  1. Immediate Audit Steps
  • Check whether ord_id contains duplicate values to confirm transaction uniqueness.
  • Verify if status_code has a mapping table or if cancelled orders (e.g. status 3) are inflating total net_amt.
  • Look for negative values or currency mixing in net_amt prior to summing revenue.
  • Confirm whether user_ref allows null values for guest checkouts.
Written by AI from this prompt. Your results will vary.

Paste your CSV header row and 5 to 10 representative rows into ChatGPT, Claude, or Gemini right after the prompt. As a follow-up, ask the model to write an SQL query or pandas snippet to run the suggested audits.

Make it repeatable

Keep using this prompt without pasting it in every time.

Keep exploring

More in Data Analysis
Data Analysis

Plan an exploratory data analysis from scratch

Generate a practical exploratory data analysis plan when starting with unfamiliar data. Frame key business questions, data hygiene checks, and visualization priorities.