Universal Data Cleaning Planner
One guided workflow for messy data: profile every column, detect problems, explain what they mean, prioritize fixes, preview the changes, apply them, and export the cleaned dataset with a data dictionary. No uploads, no installs, no code.
Your dataset never leaves the browser — safe for confidential data.
Tool interface
Analysis
Introduction
Cleaning data usually means jumping between spreadsheets, scripts, and guesswork — with no record of what changed or why. The Data Cleaning Planner turns that into one transparent pipeline: it profiles your CSV, finds quality problems, explains each one in plain English, recommends prioritized fixes, shows you the result before applying anything, and exports both the cleaned data and a data dictionary.
How it works
- Profile: every column gets an inferred type (integer, decimal, boolean, date, text), fill rate, distinct count, and samples.
- Detect: ten checks run — empty/sparse/constant columns, mixed types, duplicates, whitespace, inconsistent case, number formatting, candidate keys.
- Prioritize: findings are scored Error > Warning > Info into a 0–100 quality score.
- Fix: you choose recommendations; the engine applies them in a safe order and logs every change.
- Document: a data dictionary is generated from the cleaned dataset.
How to use
- Paste CSV (or load the example) and click Profile dataset.
- Review the Profile and Issues tabs.
- Pick fixes in the Fixes tab and click Apply fixes & preview.
- Check the change log in Preview, then export from Dictionary.
Use cases
- Auditing an export before loading it into SQL or Power BI.
- Standardizing messy CRM/marketing lists.
- Documenting a dataset for teammates.
- Teaching data-quality fundamentals with a visual, safe tool.
Best practices
- Fix whitespace and empty-cell normalization before deduplication — duplicates often hide behind stray spaces.
- Investigate “candidate key” columns before choosing a dedupe key.
- Review mixed-type columns individually rather than forcing one type.
Limitations
- Delimiters are limited to comma, semicolon, and tab.
- Very large files (hundreds of MB) may be limited by browser memory.
- Date detection covers ISO-like formats only.
FAQ
Is my data uploaded anywhere?
No. Profiling, cleaning, and documentation all run locally in your browser.
Does applying fixes change my original text?
No. Your input stays untouched; the cleaned result is produced separately in the Preview and Dictionary tabs.