Keywords: Data Journalism, Social Media, Investigative Techniques
Over the last decade, journalists have gradually expanded the range of their activities, incorporating data, code and algorithms both as tools to report with as well as tools to report on. Each year, I teach a data-oriented class at the Columbia Journalism School. Last year, we focused on the computational tools and techniques that, while not necessarily new, certainly achieved new prominence in the national election in 2016 and beyond. The vast networks of information that are created every day are simply too large for us to examine in their entirety. To get a sense of "what's on," we take feeds from algorithmic recommender systems, we scan trending topics, we focus on information shared with us by our friends or people we trust. Recently, we have seen how these tools and strategies for directing our attention can be hacked. Hence the title "Computational Propaganda." My class looked at each of these topics, in turn. Ultimately, our work produced a story appearing in the New York Times called "The Follower Factory," a long piece of data journalism that looked at what it means to purchase followers on social media platforms like Twitter.
The class is now in its second incarnation, and is part of a city-wide (New York City) effort to create new technologies, to look for new kinds of stories, that respond to this new societal condition, to the "threats to journalism" outlined above. The broad initiative is described here and you'll see that there are six campuses involved -- Cornell Tech, Columbia University, City University of New York, New York University, The New School and The Pratt Institute.
In this talk, I will use my course as a kind of case study in how data journalists examine a topic like computational propaganda.