A repeatable pre-analysis routine that catches the errors that quietly ruin reports.
Data & AnalyticsPDF · 9 pages· v1.0
4.4Dirty data is the number one cause of wrong analysis. Duplicates inflate counts, inconsistent categories split totals, hidden NULLs break averages, and a single text value in a numeric column silently corrupts a sum. This checklist is the routine you run on every dataset before you trust a single chart. It is written for analysts, data-curious operators, and anyone who works with CSVs and spreadsheets. The steps are tool-agnostic and include concrete how-tos for spreadsheets, SQL, and pandas, so you can apply them wherever your data lives. You will profile the data first (row counts, types, ranges), then work through the standard defects in order: duplicates, missing values, inconsistent categories and casing, type mismatches, outliers and impossible values, date and number formatting, and structural issues like merged headers or wide-vs-long shape. Critically, the guide teaches the discipline that separates good analysts from sloppy ones: never modify in place, always keep the raw data untouched, log every transformation, and document your assumptions so the cleaning is reproducible and defensible. The outcome: a clean, documented dataset and the confidence that your numbers are not built on sand. This is the free starter from our Data & Analytics line; pair it with the SQL and pivot guides for the full workflow.
Yes, $0. It is the starter product in our Data & Analytics line. It is complete and useful on its own; it also pairs naturally with the SQL and pivot guides.
No. Every step includes a spreadsheet method. SQL and pandas snippets are provided as a bonus for those who use them, but they are optional.
The principles apply at any size. For very large data the spreadsheet methods give way to SQL/pandas, both of which are covered. The profiling-first discipline matters most exactly when data is too big to eyeball.
Because cleaning steps interact. Deduplicating before standardizing categories can miss duplicates that differ only by casing. The guide sequences the steps so earlier fixes do not hide later ones.
Read the full refund policy and trust & safety terms.