A data scientist analyzed Donald Trump's Twitter account. Here's what he found.

(Image credit: Getty Images)

published 10 August 2016

You can tell a lot about a person from their Twitter feed, especially when you're using cold, hard data analysis. Data scientist David Robinson used "text mining" to analyze Donald Trump's Twitter account and noticed several differences between the tweets sent from an Android device (likely from Trump himself) and those composed on an iPhone (likely from his campaign team).

Namely, Robinson noted that the Android tweets were way more negative than the iPhone tweets. The most common words tweeted from the iPhone were "#makeamericagreatagain," "#trump2016," and "join," but the most common words tweeted from the Android were "badly," "crazy," and "weak."

The data also surfaced a difference in the technological savvy between the two users. "Almost all" of the tweets that included the "anachronistic behavior of 'manually retweeting'" — which Robinson described as retweeting "people by copy-pasting their tweets, then surrounding them with quotation marks" — were posted by the Android. If a tweet contained a link, a picture, or a hashtag, it was almost always the work of an iPhone user.