Data on 1.2M Facebook users from 2005 (2011) [use archive.org url in thread] | Hacker News Reader

Data on 1.2M Facebook users from 2005 (2011) [use archive.org url in thread]

michaelzimmer.org

108 pointsfeelin_googley30 comments

> to my surprise, the dataset included each user’s unique Facebook ID number.

Come on Facebook, you should be hiding your internal identifiers from third parties. Ideally each consumer of Facebook data should have its own mapping of identifiers to make it more difficult to combine datasets later on.

considering that for a long time facebook didn't have "usernames" and everyone was identified by a number this doesn't seem strange (especially as my faulty memory might mean that they created usernames post 2005)

I know they exposed identifiers in friends lists as well, long into 2013 as I recall. I remember hanging out with friends and grabbing the Inspect Source HTML from their little "Friends tile" that had a ranked list of Identifiers that were based off of how much their algorithm thought we were friends. It would display a random selection from the top 50 or so but all 50 were in there and were fairly accurate as to who I interacted with most on Facebook. My friends and I would log them all then compare notes.

I never had any illusions that Facebook wasn't spying on me: I just had illusions that they wouldn't be so stupid as to sell the only thing that makes them valuable in a raw format to customers.

Google at least just sells you targets.

One can still find the id, even when a person has chosen a username that masks it. Right click the block option of a user you are not friends with, copy the link and voila, you've got the user id now.

I mostly use that to see public but normally inaccessible pictures of people I am not friends with (i.e. pictures they've been tagged in).

Try it yourself (for some reason this only works on desktop mode): https://www.facebook.com/search/{$PUT_THAT_ID_HERE}/photos-o...

Facebook has restricted apps to only receiving app-scoped IDs since 2014.

Only took them 10 years?

That deserves a serious wat. How could you be so dumb as to release internal non-opaque IDs.

I believe we called that, "2003 engineering decisions."

It was 2005. The web was a different place back then. Although it's a bit earlier, ICQ used monotonically increasing user IDs as its addressing mechanism.

Archived versions:

2011

http://web.archive.org/web/20110217082656/http://michaelzimm...

2018

http://web.archive.org/web/20180322184527/http://www.michael...

Dataset:

https://archive.org/details/oxford-2005-facebook-matrix

cached version: http://webcache.googleusercontent.com/search?q=cache:bxG_r4W...