Much of the data Jonathon Morgan looks at these days is not for the faint of heart. Armed militias, delusional belief in a rigged election, white nationalism, Nazi symbolism, and other alarming strains of radical right thinking have advanced from the far fringes into the mainstream in this election cycle. Jonathon’s analyses examine social media’s role in this disturbing trend, and he’s recently posted some of the underlying data to data.world.
The Southern Poverty Law Center counts 276 such militia groups. A closer look at militia activity across Facebook reveals that, while many restrict their activity to closed, private communities, over 240 militia groups keep active, public Facebook pages.
This dataset contains aggregate counts of comments and interactions on 246 public pages for militia groups in the United States that have been active in 2016. It’s a subset of the data we used in recent research … into the extent to which militia group members are responding to the Trump campaign’s claims that the presidential election is rigged.
And he’s just the person for the job. Jonathon is Principal at New Knowledge (formerly Popily), where his team has built this same exploratory power right into the product, which helps data-driven teams “hypothesize, prototype, learn, communicate, and iterate with data every day.”
That’s why we’ve recently integrated New Knowledge’s exploratory visualizations into the data.world platform (more on that in a bit). First, let’s get to know Jonathon Morgan.
What’s your proudest achievement so far with New Knowledge?
Whether it’s through our automated visualization and data science platform, or our work modeling complex human behavior, we love helping people make sense of complex, confusing information, and understand new things about the world around them. For example, we’ve been proud to partner with the United Nations in informing world leaders with analytics drawn from surveys across 147 countries to help them make evidence-based recommendations for women’s economic empowerment.
What’s the next big thing on your horizon?
We’re excited about a new technology we’ve developed that captures subtle changes in human language to measure changes in people’s attitudes and perspectives. This is a powerful way to identify and understand everything from complex extremist ideologies to the aspects of a brand that inspire us as consumers.
How are exploratory visualization and analytic visualization different?
Both exploratory and analytic visualization are essential to understanding what’s valuable in your data, and communicating that value to others. Exploratory visualization is about getting the lay of the land. Does this data answer my question? Or, what questions might this data help me answer? This usually comes at the beginning of an interaction with a new dataset. Analytic visualization usually comes toward the end of the analysis process, and is about communicating something that you already know in order to share that discovery with others.
What motivates your research into the extremist far right?
Our work on the extremist far right looks at the problem from two angles. First we’re interested in understanding how these groups operate in relation to other extremist groups, like ISIS, that have successfully used social media to recruit and spread their ideology. The second is to understand how these groups have responded to the current election, and the extent to which they’ve moved from the political fringes to the mainstream.
What’s the most interesting aspect of producing your podcast Partially Derivative?
We love keeping up with all the exciting, clever, and interesting applications of data science around the world. Every week someone has invented a new approach to machine learning, or discovered a new dataset that explains something new about the world. Plus we get to chat with some of the smartest people in data science, which is amazing!
Now for the call-back to our new exploratory visualization capabilities, powered by New Knowledge.
Here’s a short video of the integration in action. For a closer look, register for data.world (it’s free) and check this out.