Home / Data cleaning
A file you can rely on.
I clean and organise Excel and CSV files merged from several sources: customer lists, products, orders, contacts. You get the data in a consistent format and a log of every change. Anything missing or doubtful stays flagged, never filled in by guessing.
What I doCleaning and organising work, on small and medium-sized files.
What I can do with your file
- Consistent formats
- Names, phone numbers, emails, dates, amounts and cities written the same way on every row.
- Duplicates
- Found using rules we agree on together (for example, the same email or the same phone number). When two rows contradict each other, I don't pick one: I flag them.
- Missing and invalid values
- Invalid emails, incomplete phone numbers, impossible dates. They stay empty and are marked in the notes column.
- Columns and structure
- Columns split correctly, clear names and the order you need for importing into another program.
- Change log
- Every change with the source row, the old value, the new value and the reason. You can check everything.
- Ready to use
- A file ready to import into your customer database, email platform, accounting software or Google Sheets, in the format required.
I don't do statistical analysis, business reporting or database migrations. I don't work with medical or sensitive financial data without a written agreement on how it's protected.
Synthetic dataThe people, emails and phone numbers are invented. The .test domains and the 0700 prefix belong to no one.
A customer list merged from three sources
A website form, an old Excel file and an address book, pasted into one file from a Romanian business. I cleaned the list with a script, using the rules below. Switch between the raw file, the result and the log.
- Namebefore: ANDREEA POPESCUafter: Andreea Popescu
- Phonebefore: 0040700123403after: +40 700 123 403
- Citybefore: cluj napocaafter: Cluj-Napoca
- Datebefore: 3 April 2025after: 2025-04-03
- Amountbefore: 1.250,00 RONafter: 1250.00
- Datebefore: 31.04.2025after: impossible date, flagged
Loading the table. If it doesn't appear, use the CSV files below.
Download the raw file (CSV) Download the result (CSV) Download the log (CSV)
The demo's rulesIn a real project we agree on them together, before cleaning.
What the script decided, and what it left to a person
It corrected
- Names: capitals and double spaces.
- Emails: lower case, trailing spaces.
- Phone numbers: one format,
+40 7xx xxx xxx. - Cities: the official name, with diacritics (“cluj napoca” becomes “Cluj-Napoca”).
- Dates: the YYYY-MM-DD format. As the source is Romanian, “12/03/2025” is read as 12 March.
- Amounts: “1.250,00 RON”, “1250 RON” and “2,150.00” become numbers with two decimals.
- Duplicates with the same email or phone and the same values: merged, with empty fields filled only from a duplicate of the same person.
It flagged, without guessing
- An email without a complete domain (
radu.stan@mail). - A phone number with only six digits.
- A date that doesn't exist: 31 April.
- A customer with no phone number and no order amount.
- Two rows with the same person and date but different amounts (1,050 and 1,250 RON). It could be a mistake or a second order. Someone who knows the customer decides.
How I workYou don't get a “fixed” file with no explanation.
From your file to a clean one
A sample
You send me a few dozen rows or the full file, with personal data replaced if you prefer.
Rules and quote
I tell you what problems I see, propose the cleaning rules, and you get the price and timeline.
Cleaning
On a copy. The original stays untouched.
Handover
The clean file, the change log and the list of rows to check.
The price depends on the number of rows, how many sources are mixed and how many rules need applying. That's why I set it after seeing a sample, not before.
Have a file you no longer trust?
Describe what it contains, where it comes from and what you need to do with it. Don't send personal data through the form; we arrange that separately.
Need a quote? Contact us with your requirements for a tailored estimate.
Ask for a data cleaning quote