does any of this shit mean anything???
Does any of this shit mean anything???
Other urls found in this thread:
github.com
twitter.com
what is their metric?
>GOOD -> positive
>FAGGOT -> negative
you braindead cunt
i'm not sure
Probably how much people speak in positive/negatibe terms.
Of course its a biais in itself
how does that bias?
i think that's exactly how this piece of shit works. it's also super fucking laggy
whats with these emojies
idk but its lagging me
Sentiment analysis
can you explain how this works? i have a small brain
People tend to voice even minor negative views much more frequently than positive ones
>podrick negative
lol?
so you're saying that it would skew more negative just based on human behavior? i guess that makes sense
fuck pod
hes neutral now
Only 30 tweets about him, even one guy having a negative opinion skews much more the "general" opinion than the other characters
this is pretty gay
ah i see. i think i get it now. pretty cool. tho that means this doesn't mean shit? it's not really what twitter is feeling about this
this is what happens when faggots are way into gay children's fairy tales
It's just a way to analyee available data, and an analyze can only be as good as the data is. People that use Twitter arent representative of the general GoT watching population anyhow, and that some characters like Podric arent not talked about means that the "general opinion" about them is meaningless
I only have a general idea of how the tool works, but it comes with a word list where every word has a score. When you give it a text it returns a score based on the sum of the matching words' score. It also takes into account a list of negations, such as "doesn't", and intensifiers.
mem, wonder what is it about D&D
>greyworm positive
>Brienne positive
lol
yeah me too. fuck em
so in essence, this doesn't mean shit?
/endthread
It means what it says it means on the picture, its the opinion of people talking about GOT on Twitter. Not every GOT watcher uses Twitter, not every person on Twitter watches GOT, not every person with a Twitter account watching GOT talks about it
depending on the model, I think phrases or groups of words can also be considered
That makes sense. At least it's a cool representation of what people on twitter think about each character? It would be tough to get the data elsewhere right?
A more accurate way to get that data would be to ask people about it directly.
First by using a general opinion poll to see who actively watches GOT to see what demographics we're dealing with, then in a second time making a second poll with an accurate representation of those population and ask them to rate their appreciation of the characters, on a scale of 0 to 10 for example.