I came across a very interesting post last week on David Robinson’s blog (varianceexplained.org) that I had to share with you.
David took some time to gather and analyze data from Donald Trump’s twitter feed.
As you probably know, Donald Trump actively and frequently tweets his thoughts and opinions as he campaigns as the Republican nominee for the US presidency this year. David’s analysis of Trump’s words revealed several trends in his Twitter feed. One interesting revelation is that Trump exhibits two distinct personality types in his posts.
The first persona is expressed through from iPhone devices, likely coming from his campaign managers. The second Twitter personality from Trump’s team is from an Android device, apparently from Trump himself.
The Data revealed a few dominant trends:
- Tweets with a picture
- Tweets with a link
- The type of devices used to tweet
- The use of hashtags. Content from iPhones seemed to be the predominant source of these.
- Content that is related to events or announcements, and includes words like “join” or tomorrow,” came almost entirely from the iPhone device
- Emotionally-charged content came mostly excluded from the Android device.
Here’s an example of analyzing “live” data that is tremendously insightful and interesting. Now, most of David’s work was done manually, using data queries with results fed into charts. But most new generation BI tools can process these comparisons extremely quickly.
I’m convinced that more data scientists and journalists should follow David’s steps and use those technologies. A lot of interesting information can come from it!
See David’s original post at https://varianceexplained.org/r/trump-tweets/
For more resources and content from David Robinson, see https://github.com/dgrtwo/dgrtwo.github.com/blob/master/_R/2016-08-09-trump-tweets.Rmd