Hacker News new | ask | show | jobs
by honkhonkpants 3609 days ago
This seems like quite an intrusion into people's private data. If certain other large companies decided to crawl all over your data for shits and giggles, I'm fairly sure HN would have its pitchforks out.
5 comments

> To answer these questions, we dug our claws into aggregate, de-identified data from a sample of about 160,000 people in the United States who shared photos of cats or dogs (or both) on Facebook.

How is looking at aggregate, de-identified data an intrusion into people's private data? ?? ?

This article is interesting, but it alludes to the level of profiling they do on their user-base, and that part is scary.
What is scary about using statistics on data your users knowingly give you?
What? This is orders of magnitude better than selling private data to ad companies, which is FB's main business. If of all things you choose to be outraged by aggregate statistics about cat vs. dog people, you may want to reconsider your priorities.
Is that what FB does? Or do they just use their data internally to target ads, without handing the data to the advertisers? I honestly don't know but I do know that people are always saying that Google sells your information to advertisers, and they don't. So I'm skeptical of this claim.

Anyway, maybe this whole analysis was based on public profile data.

In this case, I don't think the distinction is really that important; I think checking if someone is single or in a relationship to target them with an ad is more intrusive than checking if they're in a group of a few dozen thousand people who are single and like cats.
If either of these companies actually sold the data they've collected they'd hurt their own competitive advantage.
Ahah it's Facebook, what did you expect? They already sell your data and the new features like: Like/Love/HaHa Or tagging posts are just ways for them to know you better and show you relevant advertising or selling the data they have on you.
You'll probably find that they didn't use any data that's not available to 'public' viewing.

(Would be interesting to hear about their sampling methodology)

I'm not sure about that. The sentiment analysis in particular sounds like it might be using private updates.
Ready your pitchforks!

https://www.google.com/trends/