A very basic example to show the impact to your k-anonymity when you disclose your location at the state level, and then your birthdate.

Your "anonymity group" (k) is the number of people who share the exact same combination of traits as you. If k=1, you are uniquely identified.

AssumptionsAssumptions

We start with 3 assumptions to ease it up:

1. You're in the US - larger countries do better, smaller do worse.
2. You're in an averagely populated US state (MA or IN) - larger states do better, smaller states do worse
3. People live 80 years and the age pyramid is a column, so that we don't have to bother with segment size. In reality, if you're young, your group is smaller due to declining birth rates, and the same if you're old, due to deaths.
4. You're not born on Feb 29th.

CalcCalc

Total US population: 335,000,000
Number of US states: 50
Number of birth dates: 80 years * 365.25 days/year = 29,220

Here we go:

To start with, we're one out of 335M people. That's a huge anonymity set, great!

But now we disclose our state, because compliance:

335,000,000 / 50 = 6,700,000

Now we're one out of 6.7M people. That's still huge! No problem.

But now we disclose our birth date, because we live in a nanny state and we must:

6,700,000 / 29,220 = 229

Now we're one in 229 people. This is extremely fragile.

Why is this fragile you say?Why is this fragile you say?

Because every time you slip up, this gets narrowed down:

• You accidentally disclose your gender: down to 229 / 2 = 114
• Your spouse discloses the same data from the same IP address: you're done.
• You mentioned the bar you went to last night: you're done.
• You mention Bitcoin: you're done.
• Instead of state, you passed your zipcode: you're done.

PS: clean your cookies sometimes on top of never mentioning your location or birth date, to anyone.