Bots and the 2016 election Despite what we now know about the Kremlin’s coordinated efforts on social media during the 2016 election, it’s still not hard to find digital traces—or even the actual profiles—of accounts spreading eggregious examples of misinformation. In the tweet below, for example, I included a screen shot of obviously pro-Trump bot accounts—which mysteriously stopped tweeting on Nov. 2nd 2016th—that I reverse engineered by identifying accounts that had pushed a well-documented lie about Trump being endorsed by the Pope.
On Monday (Oct. 30th), the first charges were filed in the Mueller investigation on potential connections between the Trump campaign and the Kremlin. It’s major political news, so I wanted to capture as much of the relevant tweets as I could. In this post, I’ve documented the step-by-step of how I went about collecting the data. 1. Install and load rtweet Although I expect to update rtweet on CRAN very soon, the code below assumes the installed version of rtweet is at least version '0.
Twitter data gatekeeping Even before the unpresidented Twitter-centric 2016 U.S. election, researchers flocked to Twitter data (Lee, Yoon, Smith, Park, & Park, 2017; Peoples, Midway, Sackett, Lynch, & Cooney, 2016; Priem & Costello, 2010; Shuai, Pepe, & Bollen, 2012). Why? Well, aside from the obvious–that it’s an immensely popular social media platform–Twitter attracts scholarly attention because it makes a lot of its data freely available to the public via Application Program Interfaces (APIs).