Twitter track query

[aychedee]

Using the documentation for track (https://dev.twitter.com/streaming/overview/request-parameters#track) we should be able to construct a better twitter query. Current one is just "trans" and it's pretty useless and doesn't actually capture the tweets we want.

[JessFairbairn]

Is there a size limit? We should also include terms from different cultures (Kathoey, hijra etc.). We'd probably have to make sure it was in English if we're doing sentiment analysis ourselves?

[oluoluoxenfree]

We could definitely crowdsource or find a master list of different terms somewhere I'm sure.

Different cultures is genius @JessFairbairn, terrible I hadn't thought of that at all!

[RubyOffThe]

Yeah I missed that too...

[aychedee]

Cool, @JessFairbairn each 'word/phrase' is limited to 60 characters. And it's not clear from the documentation how many you're allowed to use. But I think it's quite a lot. So maybe if we set an artificial limit of 100 that would be a good start?

I've created a constant called TRACK_WORDS in hack/app.py which is joined into a comma separated list and then sent to twitter. So that would be a good place to add words or phrases. Don't worry about breaking anything.

[JessFairbairn]

Cool beans, I've added some words. Ultimately Olu's crowdsourcing idea sounds good.

I assume we're going to have old fashioned and "politically incorrect" words in, but what about out-and-out slurs? I'd imagine we should probably include everything, as people can use them out of ignorance while still being positive, and people can find different things offensive.

[aychedee]

It's up for debate isn't it? My thought is that if we want it to be accurate then it's going to need the slurs, as long as they're not overly broad. Things that are specifically transgender related rather than general gender queer Ultimately the choice of track terms is going to heavily influence the outcome... But as long as we don't change the track terms then the analysis might be interesting over time.