Data anonymization for spreadsheets: replace personal identifiers with codes and remove sensitive information
Datasets often hold sensitive personal information. Surveys, registers, staff and student records and administrative files are common data sources, and preventing personal data leaks is good practice. HILO saves you the Excel work with UNIQUE and VLOOKUP and codes all your files with the same identifiers.
HILO flags the columns with personal data, gives them a random code or deletes them, and gives you the file back in the same format:
- Nothing leaves your browser
- Free, no account needed
- The same codes across all your files
Replace ID numbers in a dataset without losing the links between files
You load one or more Excel or CSV files and review what HILO suggests, column by column. The decision is yours.
Finds the columns with personal data
ID numbers, employee or student IDs, or any code that identifies a person. Also names, emails, phone numbers, addresses and dates of birth. HILO recognizes them by the column name and by what’s inside.
The same value gets the same code in every file
Each value gets a random code. If you load enrollments and grades together, the same person has the same code in both, even if the ID is written with dots in one and without them in the other.
Generalizes what makes sense
Date of birth can become an age band, and email can become “personal” or “institutional”. A value is only derived if the dataset doesn’t already have it, and you always see the result before applying it.
Masks what’s written in comment fields
It looks for ID numbers, emails, phone numbers and the dataset’s own names inside text columns and replaces them with [ID], [EMAIL] or [NAME]. It shows you before-and-after examples to review.
Data anonymization vs. pseudonymization: what this tool does
Replacing identifiers with codes is called data pseudonymization. That’s what HILO does, and what people usually mean by “anonymizing a dataset”.
Full anonymization means no one can re-identify anyone, not even with outside information. No tool can guarantee that on its own. HILO shows you what’s left and helps you decide.
Uncommon combinations: a quick check before sharing
HILO counts how many people have a unique combination of the remaining columns. It matters mostly if the dataset will be published or shared outside your team.
HILO suggests which column to generalize or remove, and you apply it with one click if it helps. For example, in a fictitious dataset of 400 enrollments, 110 people had a unique combination of age, sex, town and program; removing the town brought it down to 19.
What it doesn’t do
Better to know up front.
It doesn’t work on text documents
It processes spreadsheets: Excel and CSV. For interviews or reports in Word or PDF, you can use HILO’s text tools.
It can’t catch everything typed by hand
Names that aren’t in the dataset, nicknames or details that give someone away can stay in comment fields. If that text will be shared, read it first.
Nothing leaves your browser
The dataset is read and processed on your computer. It isn’t uploaded to HILO or to the AI, it doesn’t spend hilos and no account is needed. The codes live in the tab while it’s open: to keep them, you download the lookup table.
It’s pseudonymization: identifiers are replaced with codes or deleted. People often call it “anonymizing a dataset”, but other columns can still single a person out. HILO shows you how many uncommon combinations are left and what lowers them.
Replacing the data that identifies a person (ID numbers, employee or student IDs, names or any code that identifies them) with a code, so they can only be re-identified with information kept separately, such as the lookup table.
It helps, but it doesn’t replace a case-by-case assessment. GDPR defines pseudonymisation in Article 4(5) and treats pseudonymised data as personal data. What the law requires for your dataset depends on why and with whom you share it. HILO doesn’t provide legal advice.
Only if you downloaded the lookup table (original → code), which is off by default. Without it, not even HILO can reverse the codes: they live in your browser while the tab is open.
Yes. If you load several files together, the same ID gets the same code in all of them, even if it’s written differently. For a second wave or the next year, you load the previous lookup table and the codes are kept.
Yes, and no account is needed. It doesn’t use AI or spend hilos.
Excel (.xlsx and .xls) and CSV. The file comes out in the same format, with the other columns untouched. The Excel file comes out clean: one sheet, no formatting or file properties.
Remove identifiers before you share your dataset
Load your Excel or CSV file and within minutes you have the dataset without identifiers, a record of what was done and, if you ask for it, the lookup table.