Data Anonymiser
For each column choose to keep it, remove it, blank it, mask it, shorten it, replace it with repeatable codes or labels, or round it. Likely columns are set for you.
Drag a file here or choose one
Worked examples
Masking a row
Name,Email,Phone,Salary Ann Lee,[email protected],07700 900123,31250
Name,Email,Phone,Salary Person 001,a***@e***.com,***** **0123,31000
Name as a label, email hidden, only the last 4 digits of the phone number shown, salary rounded to the nearest 1,000.
About Data Anonymiser
Test data, training material and examples for suppliers often begin life as real customer lists. Before such a file is shared, the personal details need to go. For each column you choose what to do: keep it, remove it, empty it, replace every letter and digit with a star, show only the last few characters, hide most of an email address while keeping its shape, or keep only the year or month of a date.
Labels such as Person 001 replace each different value with the same label every time, so counts and groupings still work without the real names. Repeatable codes do the same with a code built from the value and a text you supply. Shuffling moves the values of a column between rows, which keeps the overall figures but breaks the link to the person, and rounding makes amounts such as salaries less exact. Columns that look like names, emails and phone numbers are set up for you.
How to use Data Anonymiser
- Paste your CSV or open a file.
- Look at the choice made for each column and change it where you want.
- Fill in the extra setting some choices ask for, such as how many characters to show.
- Check the result carefully, then download it.
Common uses
- Creating test data from a customer export
- Sharing a sample with a supplier or a trainer
- Hiding names and contact details before analysis
- Reducing the detail in dates and amounts
Frequently asked questions
Is this enough to make data anonymous?
It is a helpful step, not a guarantee. A combination of postcode, date of birth and job title can identify someone even with the name removed. Check what is left, remove what you do not need, and follow your organisation's rules.
What is the difference between labels and codes?
Labels count up in the order values first appear, such as Person 001, Person 002. Codes are built from the value and the text you give, so the same value gives the same code every time, even in another file made with the same text.
Can a code be turned back into the original?
Not directly, because the original is not stored in it. But anyone who has a list of possible values and your text could try them all, so treat codes as a way to hide values, not as encryption.
How does shuffling work?
The values of the column are rearranged between rows using a seed. The same seed gives the same shuffle. The overall figures, such as the total and the average, stay the same.
What does the email option do?
[email protected] becomes a***@e***.com. The first letter of the name and the first letter of the domain are kept, along with the ending, so the shape is still recognisable.