absolutely. running for 100,000 tweets rn. give me a couple minutes to upload

that sample size small as hell op

Found HipHopByTheNumbers account

most recent 650 tweets, running with bigger size now

fuc gotta make an alt now

op said

dont make me upload the full code for you guys

smh this lightwork

Nah I believe you OP I was just laughing at Danny

This is how you’re spending your Sunday

I'd imagine that slang or text that needs human interpretation can lead to inaccurate results. Is there any way to account for that? For instance, if a tweet says "wlr is sick", that would be a negative tweet if we go by the real definition, while it really is a positive one.

Literally nobody says the word “sick” to be negative lol go outside

I've never done any sentiment a***ysis, but I assume the libraries use to perform the a***ysis already have some sort of dictionaries built in that detect the words and assign them as positive/negative. You really think the programmer has build that with urban slang in mind lol?

In my previous post I said "sick" because it was the first thing that came to my mind, but imagine any slang word in the hip hop world that could be interpreted as ambiguous if we go by a book definition.

this fye

how would you even calculate this lmao

good point. Im making sure to account for slang like fire and sick when making the predictions.

Look up the textblob library. It is fire also adding emphasis / polarity to common words like fire, trash, sick etc.

in on classic

make a spreadsheet bro

dope, i'll check it out. I'm into data a***ysis but i've yet to get into sentiment a***ysis.

How insecure you must be to make a thread like this.