Thread by @conspirator0

We monitored #DemDebate (also #DemocraticDebate/#DemDebate10) traffic for the 10th Democratic presidential debate, as we've done with previous debates. Traffic peaked just shy of 8000 tweets per minute, higher volume than any debate other then the 9th.

cc: @ZellaQuixote

@ZellaQuixote The most frequently discussed candidates were @BernieSanders, @MikeBloomberg, and @ewarren. Like the last debate, the right-wing presence in the network is fairly small, mostly @TrumpWarRoom tweeting about @JoeBiden and @DonaldJTrumpJr tweeting about @MikeBloomberg.

@ZellaQuixote @BernieSanders @MikeBloomberg @ewarren @TrumpWarRoom @JoeBiden @DonaldJTrumpJr Sentiment analysis for tweets mentioning each candidate, before, during, and after the debate. Warren and Sanders entered the debate at the low end of the pack and emerged with the highest sentiment scores. Gabbard and Bloomberg wound up with the lowest.

@ZellaQuixote @BernieSanders @MikeBloomberg @ewarren @TrumpWarRoom @JoeBiden @DonaldJTrumpJr What are the negative tweets about, anyway? We pulled recent tweets for accounts whose #DemDebate tweets for the 7th through 10th debates all had low sentiment scores (<0.1), and filtered those result to tweets mentioning the candidates with negative sentiment scores.

@ZellaQuixote @BernieSanders @MikeBloomberg @ewarren @TrumpWarRoom @JoeBiden @DonaldJTrumpJr Here are wordclouds for the negative tweets mentioning each candidate. "Trump" is unsurprisingly a dominant theme in the negative tweets mentioning every candidate. Interesting candidate-specific terms are included in the figure.

@ZellaQuixote @BernieSanders @MikeBloomberg @ewarren @TrumpWarRoom @JoeBiden @DonaldJTrumpJr Sentiment score used in this analysis is the combined score from the VADER algorithm, described in detail in this paper:
comp.social.gatech.edu/papers/icwsm14…