Count Love Project Reveals Protest Patterns

BU student’s website tracks all demonstrations against current administration

Test-17-1752-PROTESTS-071Earlier this year Tommy Leung (left) and Nathan Perkins (MED’18) launched a website that chronicles all protests against the current administration, hoping to give the data to government officials as evidence of the issues their constituents are passionate about. Photo by Jackie Ricciardi.

Meet Tommy Leung and Nathan Perkins. Perkins (MED’18), a PhD candidate in neuroscience, who studies songbirds, and Leung, a software engineer with a PhD in engineering systems from MIT, have a side gig together: counting protests.

Since the 2016 presidential election, Americans have been protesting issues from health care to immigration. Earlier this year, Leung and Perkins launched their site Count Love, which chronicles all protests against the current administration.

They wanted “to turn that protest activity into something that could be used to make a larger point,” Perkins explains. Leung says they hoped their data could eventually be given to government officials as evidence of the issues their constituents are passionate about.

To gather strong, reliable data, Perkins and Leung opted for a conservative approach: they skipped Twitter and Facebook in favor of local media accounts of American protests. Since January 20, 2017, they have recorded about 4,296 protests with more than 5,402,011 attendees (as of October 11, 2017), numbers they consider likely to be an underestimate.

Some of the most recent spate of nationwide protests occurred following the violent confrontation between white nationalists and counterprotesters in Charlottesville, Va., on August 12, 2017 (below).

Map showing the protest statistics on size and location of American protests on the issue of the Charlottesville Confederate Civil War monument. Source data from the Count Love website.

Charlottesville Protests

So far, Perkins and Leung have confirmed 351 Charlottesville-related protests involving 64,198 people, and while many have occurred in the Northeast, the two point out that “Charlottesville protests, like protests about health care, immigration, and civil rights, appear all throughout the country.” For them, the national distribution of the Charlottesville protests—and others—emphasizes that people all around the country care about the same issues. “We are not as different or divided as is often portrayed in the national media,” Perkins says.

He and Leung started their project after attending the Boston chapter of the Women’s March on January 21, 2017. “It was nice to see that so many people came out,” says Perkins. “But we were walking back and it was like, well, did that accomplish anything?”

They started by overlaying all the women’s marches onto a map of the United States (below), “just to say, look, people from all parts of the country came out,” says Perkins. “Just something a little more permanent.”

They used data from the Crowd Counting Consortium cross-referenced with reports from local newspaper and television outlets to document where protests occurred and how many people attended. According to the data they gathered, there were 425 total marches involving 4,124,543 people throughout the United States.

Map showing the size and location of Women's Marches, American protests in support of women and women's rights. Source data from the Count Love website

Women’s Marches

Soon after the Women’s March, people began protesting the Trump travel ban, and then other issues. “We realized there was this momentum,” says Perkins, adding that while big protests in cities like New York get the most national coverage, there are many more smaller protests happening in cities and towns nationwide.

To capture the full scope, Perkins and Leung built a software tool that visits local online newspapers and television station sites every night and identifies articles that include the terms “protest,” “rally,” or “march.” They parse the articles manually and pull out relevant data, including when and where a protest occurred, what was being protested, and how many people attended. The project started with a few hundred local newspaper and television sites, but now includes more than 2,000. They update the site nightly with the most current protest data. Below are their results through October 11, 2017.

Bar chart showing weekly American protests by attendance count and issue. Source data from the Count Love website

Weekly Protests by Attendance Counts

Unlike the Women’s March, which was a nationally organized series of protests on a specific day, the American protests over immigration initially emerged after the first executive travel ban and continued over a longer period of time.

“The travel ban was much more organic; people just went to their local airports,” Perkins explains. “There were some very big and sustained protests at cities that do act as crucial ports of entry, but the travel ban mobilized protesters in large numbers throughout the country,” as seen in the map below. Perkins adds that protests occurred in places that aren’t particularly associated with large immigrant populations, such as Boise, Idaho, Joplin, Mo., and Portland, Maine. In the month following the initial travel ban on January 27, 2017, Perkins and Leung confirmed 433 American protests related to immigration with at least 219,045 attendees.

Map showing the protest statistics on size and location of American protests against Donald Trump's travel ban executive order. Source data from the Count Love website

Travel Ban Protests

They also mapped the most commonly protested issue in each state (below). Immigration is protested the most in southern border states like California, Arizona, and Texas, and in states with large immigrant populations like New York and Massachusetts. Civil rights protests are most common in central states such as Colorado and Minnesota, while racial injustice tops the list in southern states such as Virginia, Georgia, and Tennessee. Only in Nebraska are executive protests—general protests of the current administration not associated with a specific policy or event—most common.

“People will put the effort into protesting and standing up for an issue if it has more immediate relevance,” Perkins says. “Ten people come out in this teeny town, and they’ll hold a little sit-in or talk about how much they appreciate the Affordable Care Act.”

Protest statistics map showing the most commonly protested issues in each state. Source data from the Count Love website

Most Commonly Protested Issues by State

After hearing so much about the divide between urban and rural, Perkins and Leung did their own analysis. They broke down the issues people are protesting in urban and rural counties, defining urban as more than 200 people per square kilometer and rural as less than 200 people per square kilometer. They discovered that people are generally protesting similar issues, including immigration, civil rights, and racial injustice, as shown in the pie charts below.

“For the most part, there was surprising consistency in their priorities, which was kind of nice to see,” Perkins said. “It suggests that a lot of values are held throughout the country.”

“We aren’t all that different,” Leung adds.

Pie charts showing the most commonly protested issues at American protests in rural and urban counties. Source data from the Count Love website

Most Commonly Protested Issues in Rural and Urban Counties

Some highly populated states like California and New York, which have a lot of protesters overall, also have a lot of protesters relative to population, as seen on the bubble map below. By contrast, other large states like Texas and Florida, which have a large number of overall protesters, do not have a lot of protest activity relative to population. There are also states with small populations, like Maine and Vermont, that have a lot of protesters relative to population. Massachusetts falls in the middle population-wise, but has had a lot of protest activity.

Cartogram bubble map showing American protest statistics on the number of protesters relative to the population of each state. Source data from the Count Love website

Number of Protesters Relative to State Population

Micah Sifry, a cofounder of Personal Democracy Media, was so impressed by Count Love that he invited Perkins and Leung to speak at the 2017 Personal Democracy Forum. “They are a great example of people with tech skills and data smarts taking on a hard problem and solving it in a fresh way,” Sifry says. “It’s very hard to get an accurate picture of political protest activity and no method will be perfect, but their approach—which is very conservative—definitely offers a solid baseline.”

Zachary Steinert-Threlkeld, a political scientist at UCLA who uses Twitter geotags to record crowd size at protests, thinks Count Love is “really promising.” He points out that measuring American protests is a useful way of determining how citizens feel about the many policy issues that come up between voting opportunities. “The protest movement in the United States appears to be well sustained and broadly based both geographically and substantively,” he says.

Perkins and Leung are developing tools to more directly capture crowd size at protests while protecting individual privacy. They built an iPhone app—also called Count Lovethat allows users to indicate they attended a protest (and verifies using GPS location), yet ensures the information cannot be traced back to a specific phone.

They also hope that someday their approach of extracting structured information from local newspaper and television stories can be adapted to look at topics like gentrification, housing, and education on a national scale.

“We hope that Count Love will serve as a resource for future research in political engagement and social activism,” Perkins says. “[Protest] has played a crucial role through so many generations of reform in our country, and by recording those acts of dissent, we can do a small part to amplify and remember that activism.”

Time-lapse video of American protests since January 2017.

BU Today article by Catherine Caruso. Data Visualizations and video by Count Love.