← All articles

How to anonymize music school spreadsheets for a tuition comparison

To prepare music school spreadsheets for a tuition comparison, work on copies, replace identifying references with consistent arbitrary IDs, and remove personal details from every sheet and field. Keep only the calculation evidence you need. Renaming families alone does not establish anonymity: exact dates, unusual notes, codes and hidden workbook content can still identify people.

Use the same arbitrary family ID in every file

Assign neutral codes unrelated to names, initials, contact details or existing personal identifiers. Use the identical family code in attendance, payments and opening balances.

The following structure is entirely fictional:

File Shared family ID Calculation fields
Attendance FAM-910 Lesson date, status, charge, required cancellation evidence
Payments FAM-910 Payment date, amount, optional neutral receipt code
Opening balances FAM-910 As-of date, amount, owed-or-credit direction

Changing the ID separately in each file breaks the relationship. Using a hash of an email address is not the same as removing the identifying relationship, either.

Keep any private lookup between neutral IDs and people in your own secure school records, outside the copies you share. A retained lookup means the school can still reconnect the records to people; it is one reason replacing names alone should not be described as guaranteed anonymization.

How do you anonymize music school spreadsheets without breaking the calculation?

Separate identifying information from mathematical and policy evidence. Remove names and contact details; retain the dates, amounts, outcomes and notice evidence needed to test the balance only if the resulting copy is genuinely safe to share.

Keep when needed for the comparison Remove or replace
Consistent arbitrary family ID Names, initials, emails, phone numbers, addresses
Lesson and payment dates Free-text bank descriptions identifying the payer
Numeric charges, payments and opening amount Family notes and unrelated personal details
Attendance outcome and cancellation initiator category Teacher names and identifying narrative
Required notice hours or timestamps Unnecessary identifiers in comments, formulas and filenames

Don’t change a cancellation timestamp to make it qualify for credit. Don’t round amounts simply to obscure them and then claim that the result verifies the original ledger.

If preserving the necessary dates or amounts leaves a person identifiable in context, stop. Use a fully fictional example instead. A meaningful balance comparison is not a reason to share an unsafe copy.

Inspect the workbook beyond the visible rows

Hidden columns and extra worksheets are still part of an XLSX file. Check every tab, hidden row and column, comment, formula, embedded item, external link and filename. Remove unnecessary content rather than hiding it.

Microsoft’s Document Inspector guidance explains how to find hidden data and personal information in supported Office files. Run inspection on a copy; removing hidden content may affect a workbook. Inspection is a useful check, not proof that remaining rows cannot identify people.

For a values-only comparison file, verify that pasted values haven’t retained identifying text and that IDs, dates and currency amounts remain correctly typed. If you export CSV, reopen the exported file itself and inspect all rows. Checking the workbook you exported from is not the same as checking the output.

Test that the three files still connect

Before sharing anything, compare the family-ID sets across attendance, payments and opening balances. Investigate missing and mismatched IDs rather than treating a missing opening as zero.

Check these controls locally:

  1. The copied payment total matches the intended source period.
  2. The opening balances use the agreed period-start cutoff.
  3. Each required cancellation-evidence field is present.
  4. The copied adjustments and closing totals match your owner-approved comparison, without undocumented changes.
  5. The exports contain no other families’ identifying notes or unused tabs.

A matching total does not validate anonymity. Calculation checks and privacy checks answer different questions.

Use a fictional set when you only need to test headings

Tufton’s preparation page provides three matching fictional CSV templates and a heading-only mapping check. The public check is for column headings, not family rows, amounts or personal identifiers. Its checklist does not inspect files or guarantee anonymity.

For a first look at the explanation rather than your data, try the public sample: it uses fictional records and shows the source of a balance movement. Signed-in comparison is for fictional or fully anonymized files; real personal-data uploads are not enabled. If you are uncertain whether a copy is anonymous, don’t upload it.