Using Python to find Russian Twitter troll tweets aimed at Canada

by Roberto Rocha

pandas regex text mining

Paid internet trolls linked to the Russian government tried to influence the U.S. election with divisive social media posts. But they also targeted Canadians, at a smaller scale. Here's how I used Python to find tweets aimed at dividing Canadians among a trove of 3 million troll tweets.

It's known that the Internet Research Agency, a troll farm in St. Petersburg, tried to sow division among Americans by flooding social media with inflammatory content on both sides of a debate. A trove of 3 million tweets associated with these troll accounts was made public. Using pandas and some regex, I found about 8,000 that were seemingly aimed at polarizing Canadians. This talk will briefly go over this research that was made into a news story for the CBC.