In front of the cameras, election night at a broadcast network is smooth, polished and full of gigantic maps. Results come in and go onto the screens like clockwork.
Behind the scenes at NBC News again this year, Penn undergraduates worked feverishly to support the work of the network’s analysts and producers by tracking voter turnout, turning granular results into usable data, and building graphics for instant use on viewers’ screens.
The opportunity is part of the Penn Program on Opinion Research and Election Studies (PORES). The more than three dozen students had spent months testing and simulating how they would work with the data, going back to the spring primary elections.
For some, it was the culmination of four years of study and practice. “It’s been like the journey of my undergrad experience,” says Justine Orgel, a fourth-year majoring in political science from San Francisco. “I came to Penn knowing I wanted to do this.”
Penn is the only university in the country to have a long-term relationship like this, says John Lapinski, Robert A. Fox Leadership Professor of Political Science and elections unit director for NBC News. Lapinski established the PORES partnership with NBC News in 2014.
“The students working with PORES play a critical role in our team’s projections and calls on election night. This is a fantastic learning opportunity for them to see the work that goes into the decisions behind the scenes at NBC News,” Lapinski says. “These types of hands-on experiences give students a very strong background in data analysis and other topics. It’s something they can’t get anywhere else; we’re the only university that has this type of long-term relationship.”
Seamus O’Brien, a fourth-year from Washington, D.C., majoring in philosophy, politics, and economics and minoring in data science, has been working with PORES for three years.
“It’s honestly mesmerizing,” he says of being an integral part of the team in the room where races are called. “In the past, I was stationed in the same room very close to where all of the professors were and witnessed Dr. Lapinski call in the calls, and it was a really exciting experience.”
This year, he worked on the precinct team, helping to transfer local results to NBC’s analysts in a format they could use, turning raw data into easily usable material.
“Every county reports their votes a little bit differently,” says Charlotte Davey, a third-year from Wilmington, Delaware, majoring in mathematics. “Some use a PDF, some use Excel, and sometimes they report them using a map.”
The PORES team sorted the data so they were given to the analysts in a standardized format. O’Brien says he worked for months prior to Election Day writing code specifically to “scrape” precinct-level data from PDFs posted by local elections officials, and he then pulled the data into a modeling database. In some cases, where officials had scanned PDFs off-center or slightly turned, they had to be entered manually, he says.
Jackie Balanovsky, a third-year from Las Vegas, majoring in psychology and philosophy, politics, and economics, worked on Election Day calling local officials in Michigan and Wisconsin to track voter turnout, an important part of elections projections.
“As people were actively voting, we were checking in throughout the day at regular times to get updates, to see how things were changing over time,” Balanovsky says.
She’s has been involved with PORES since the summer after her first year, moving into decision desk work last fall, and says the experience taught her that running elections is a lot more complicated than most people realize. “This gave me insight into the intricacies of election administration and a better understanding of this complex system,” she says.
Davey was on the precinct reporting team for Pennsylvania, monitoring 16 counties as results came in. “We kept tabs open on our computers and just flipped through them and kept hitting refresh until we got something,” she says.
Around 10 p.m. results started coming in, very quickly in the case of some of her teammates. The group helped each other if results were difficult to scrape using software or if the volume was overwhelming. “I was reading out vote counts so they could type them up because some of them we can’t sort with code,” Davey says.
She started working with PORES over the summer and says that as a math major she likes to see how data and mathematics have practical application in areas of interest to her. “To get to be a part of reporting the data in an election that was watched by the whole world was super cool,” Davey says.
Orgel was assigned to the exit polling team, adding to the understanding and analysis of how and why people voted.
Exit polling data is collected throughout the day but can’t be released until the polls close. “Once they do, that’s where the magic happens,” Orgel says.
Her role was to build graphics to report the exit poll results that were then used by NBC producers on the air. Much of that work was done weeks and months prior, creating hundreds of graphics that results could be quickly entered into to help viewers understand the election’s dynamics. On election night, she also built special graphics requested by the NBC News team to drill down into specific topics.
“We had to get ready and think about the best way to visualize these things—whether it’s a pie chart, a table, or a bar chart—and send them out to producers,” Orgel says. They had to carefully consider both colors and word choice: “It needed to be digestible for the anchors as well as the viewers.
“To see them on live television and see the anchor walking through what these results mean was pretty exciting,” she says.
