Twitter-Mining

Acquisition and visualization of tweets.

View the Project on GitHub saverymax/Twitter-Mining

For an hour during the evening of the FBI raid of Michael Cohen’s office, I streamed some tweets. I collected any tweets including the word Trump, guns, nda, Cohen, Stormy, or FBI.

More or less, this is what people were talking about

The most commonly used words. Most common words on twitter during FBI raid on Michael Cohen's office

And the most common hashtags Most common hashtags on twitter during FBI raid on Michael Cohen's office

Word network of words used together at least 100 times Most common coocuring words on twitter during FBI raid

And to thin out the network a little, word network of words used together at least 200 times Most common coocuring words on twitter during FBI raid

To see what the tweets were actually about, I ran a topic model, using LDA. The words in the legend are those that best represent the topics. LDA is probabilistic, and I have included only tweets that have a greater than 80 % chance of being in the assigned topic. Move your mouse over each point to see the tweet it represents.
See the plot here

It’s interesting to see the topics, but it’s more interesting to see the coocuring word networks for each topic.

offices, daniels, michael, man

Most common coocuring words on twitter during FBI raid

ed, like, news, michael

Most common coocuring words on twitter during FBI raid

doj, hilary, clinton, loudobbs

Most common coocuring words on twitter during FBI raid

tell, let, right, act, ag

Most common coocuring words on twitter during FBI raid

president, potus, make, sessions

Most common coocuring words on twitter during FBI raid

So that’s what twitter looked like, more or less, during the evening after Michael Cohen’s office was raided by the FBI