Item — HN Reader

Depends how many people in the same post/zip code have the same birth month, are male/female, etc.

E.g. If there’s 1 female in post/zip code XXXX born in 09/79, doesn’t matter if you’re only including the month. There’s a good chance you can work out who she is.

It’s a sliding scale dependent on the other attributes in the dataset, and the number of distinct values across all the attributes.

The only way to make data completely anonymous is to remove things like birth date tbh.

I'm no expert, but I think removing the postal code is the obvious way to anonymize.

If you can't restrict to a specific geographic area, then who knows who you are.

Pragmatically speaking, I noticed that in their "test toggling variables" widget at the end, removing postal code instantly dropped my probability of detection to 0, no matter what other fields were active.

So perhaps your geographic location is the information you should be most paranoid about sharing.