A data cleaning tool that shows you every change it made
Automatic cleaning is easy to build and hard to trust. A tool that quietly rewrites values is worse than a messy file, because the mess is at least visible. DataMadeClean takes a CSV or Excel file, fixes what can be fixed unambiguously, flags what cannot, and lists every change so you can see exactly what happened to your data.
No account needed · up to 500 rows free · every change listed
Most of the risk is in the fixes, not the mess
The failures that cost you are the ones you cannot see.
- A tool that standardizes a date column by guessing day/month order, and gets it backwards for half the year
- A postcode read as a number, so 0800 becomes 800 and matches nothing afterwards
- A near-miss name treated as the same customer and silently merged away
- A blank filled in with something plausible, which is now indistinguishable from real data
- A cleaned file with no record of what changed, so nothing can be checked or undone
What it changes
Fixes it will make on its own
These are unambiguous: there is one correct answer and the data supplies it. Every one is listed in the summary.
| What | Before | After |
|---|---|---|
| Whitespace and invisible characters | john smith | john smith |
| Personal name casing | JOHN SMITH | John Smith |
| Email addresses | JOHN@CO.COM | john@co.com |
| Phone numbers | 0412987654 | 0412 987 654 |
| Dates, to the column's own format | 2026-04-03 | 03/04/2026 |
| Numbers and currency | $1,204.50 | 1204.5 |
| Yes/no columns | yes | Y |
| State abbreviations | qld | QLD |
| Duplicate rows | the same row twice | first kept, rest removed |
Duplicates are compared after whitespace and casing are normalized, so a row hiding behind different capitalization is still caught.
What it flags
Things it reports instead of guessing at
Each of these is left exactly as it is in the file, and raised for you to decide.
| What | Example | Why it is not fixed |
|---|---|---|
| Two-digit year | 31/12/25 | 1925 or 2025 — the file does not say |
| Value that is not a valid phone number | 04129876 | Too short; padding it would invent digits |
| Email domain that looks misspelled | ana@gmial.com | Suggests gmail.com, but sending to a guessed address is worse than asking |
| Blank in a column that is usually filledFree | an empty cell | Reported as “filled in for 91.7% of rows — this one is empty” |
| Suburb, state and postcode that disagreePro | Sydney, QLD, 4000 | Which of the three is wrong is unknowable from the row |
| Unusual value in a numeric columnPro | 98500 among values of 45–68 | Reported with the range it fell outside |
| Near-duplicate contactsPro | two spellings of one name | Merging the wrong two records is unrecoverable |
What it leaves alone
Deliberate non-changes
Restraint is most of what makes an automatic pass safe to run.
- Zero-padded identifiers. A column holding 00123 is read as text, so customer IDs, product codes, BSBs and postcodes keep their padding — and padding is never added to a value that did not have it.
- Genuinely numeric columns. Quantities, prices and ratios are still handled as numbers, so the checks that depend on that keep working.
- Dates already stored as dates. A real date cell in an Excel file stays a real date cell. Only text that looks like a date is rewritten.
- A column that is already consistent. Suburb names written entirely in capitals stay in capitals; the column's own style is followed rather than replaced.
- Date columns that are not day-first. A column proved month-first by a value like 04/13/2026 keeps that order rather than having a regional convention imposed on it.
- Service numbers. 1300, 1800 and 13 numbers keep their own spacing instead of being grouped like a mobile.
Questions about the cleaning tool
Does it change my file without telling me?
No. Every change is listed individually — the row, the column, the value before and the value after — so the cleaned file is auditable rather than something you have to take on trust.
What does it do when it is not sure?
It flags the row and leaves the value alone. A two-digit year, a number that is not a valid phone number, a suburb and postcode that do not agree: all of these are reported for a person to decide rather than guessed at, because a confident wrong fix is worse than an obvious mess.
Do I need an account to try it?
No. Files up to 500 rows are free and need no account. An anonymous upload and its result are deleted an hour after cleaning.
What is included on the free tier?
Whitespace, casing, names, emails, phone numbers, dates, numbers, yes/no columns, duplicate rows and unexpected blanks are all handled on every plan. Address validation against G-NAF, outlier detection, near-duplicate matching, imputation, cross-field checks and date sanity checks are Pro features.
Upload a file and see every change before you download it.
No signup required.
Clean a file free →Get in touch if you have a question first.