Open OLI9292 opened 7 years ago
from Thomas,
Distribution of Twitter's personality
Something dedicated to filtering the news
We can get any public data from places like Twitter and Reddit. For Twitter we can use the search endpoint to look up Tweets made from / to specific accounts, or we can use the streaming endpoint, which I was using before, and that gets Tweets in real-time that match keywords you pass to it in an array.
I like both of your ideas. I'm digging the idea of making a visualization that's updating on its own, whether in real-time with the streaming API, or we can poll Twitter's main API. It might be cool to use a bubble chart similar to this one and track a few dominant moods that Twitter is currently in... We could perhaps filter by type of news story as well.
One thought I'm having is that we should just make this about United States state of affairs / public mood, because with world politics we have to deal with languages. Lmk what you think.
My primary concern with the current dataset (i.e. the data in
fetched_tweets.txt
) is that 90% of the Tweets concern Donald Trump. So, it seems like any analysis / visualizations we do will be primarily focused on Trump.Other ideas would be to:
Track major media outlets on Twitter and conduct NLP analysis on their content. We could pick out names from their Tweets. Trump will probably still dominate this data.
Make a Donald Trump focused analysis.