The most interesting thing about the "Thanksgiving Effect" study is what it tells us about the limits of data anonymization

Originally published at:


Technically about Christmas dinner, but it seems on point.

This is us.


My turkey was gerrymandered and I got the whole tail and neck!

I learned a long time ago, before mobile phone tracking, that finding out the true name of an individual from a supposedly anonymised dataset is surprisingly easy. There are just certain required pieces of data (gender, age range, ZIP code) that immediately narrows candidates down to a manageable number. That’s before you get into things like text analysis and other clues dropped inadvertantly by someone trying to hide his identity.

Fortunately the data brokers are also slipshod and lazy, which means there’s a lot of polluted data floating around about people, myself included.


In short: It’s astonishing what you can infer about personal relationships from just a few datasets.

Don’t worry. We were told not to worry.


I’m happy to say I turn my phone off outside of business hours.


I’ve got to stop using mine as an alarm clock now. :sob:

You could just put it into aeroplane mode, and that’ll help the battery life as well as not communicating your location.
On the other hand, if a phone ‘turns off’ at an address at (say) midnight, and turns back on in the same location in the morning, there’s still a very good chance that the phone belongs to someone sleeping at that location.

Basically what we should be doing is giving our phones to our cats at night, so our location data gets nicely muddled overnight. (Assuming your cats are like mine, and sleep all day and do their own thing at night)

1 Like

Of course, I had to get the cat that is incapable of sleeping more than two feet away from me. :crying_cat_face:

This topic was automatically closed after 5 days. New replies are no longer allowed.