Sept. 2, 2020

---

Elecurator project team evaluating political landscape through a big data lens

MILWAUKEE – The Northwestern Mutual Data Science Institute is connecting students to the political landscape in a nontraditional way – through a funded research project that uses data science to shed light on what might be driving voter behavior during the 2020 presidential election campaign. The Elecurator project, which began in January 2020, uses a variety of data sources, including online and social media, traditional polling methods and political advertisements, to determine what issues are on the minds of voters.

The project is led by Co-Principal Investigators, Dr. Purush Papatla, Co-Director, Northwestern Mutual Data Science Institute and Professor of Marketing at UW-Milwaukee, and Dr. Amber Wichowsky, Associate Professor in the Department of Political Science at Marquette University and Director of the Marquette Democracy Lab. The project team includes a mix of undergraduate and graduate students in various degree programs, with varying experience and knowledge of data science and politics.

The Democratic National Convention in Milwaukee provided the project team a valuable opportunity to evaluate the political landscape through a big data lens. “We can use data science to investigate Americans’ priorities surrounding the election, including how voters and candidates are talking about current issues,” said Wichowsky. “The students are looking at the issue landscape of the 2020 election, in particular at how the three crises facing the country (healthcare, the economy and racial justice) are reshaping what Americans are thinking about and prioritizing as we head into the November election.”

Prior to the DNC, and especially in the weeks leading up to it, the students collected and analyzed social media data about the elections to look at the issues garnering the most attention online. They coded candidate speeches, debate transcripts, and political advertising to look at the issues the candidates are prioritizing in their campaign messages. During the DNC, the students used a method called topic modeling to understand what topics are highlighted during speeches and how the general electorate are responding digitally.

“This project is perfectly aligned with the mission and vision of the Northwestern Mutual Data Science Institute,” said Papatla. “We’re providing opportunities for students to cultivate the skills they will need for a future in data science while introducing them to societal and political issues that have an impact locally and nationally.”

In fact, three of the former team members have already obtained full-time data science positions in companies around Milwaukee and the remaining students are pursuing additional research opportunities at their universities. Ben Garski, a recent UWM graduate and former team member, just started a job at Northwestern Mutual. “Being able to say I did natural language processing, a newer machine-learning process, as part of this project really makes you pretty employable.”

Jake Beihoff, another recent UWM graduate, decided to stay on the team even after he got a full-time job with MolsonCoors. “Working with topic modeling and text analytics during this project opened my eyes to how much I enjoy doing this. You want to understand what people are saying and what the topics of conversation are, but you don’t necessarily have a pre-defined idea of those topics. I like that because it really lets the data speak to you; it’s an endless cycle of problem solving. I’ve found that this experience translates easily into my work because understanding how consumers engage with your brand is so important. Analyzing digital conversations is one way of doing that.”

Savannah Charles, a senior at Marquette, used her experience on the project to obtain an Honors Research Fellowship. “Since I didn’t have any prior background or coursework in research methods, I’m grateful that this project allowed me to explore strengths I didn’t know I had. I didn’t have an appetite for statistics prior to this but I really enjoyed working with the math and data side of things.”

Charles is analyzing the issues that divide the electorate along generational lines or which issues might engage younger generations versus older generations. It’s just one way to introduce a whole new generation to data science, demystifying the field for those students who might not have an understanding of how data impacts all areas of our world or who feel like they might not have the skills to advance in this space. “Using the tools and seeing how they are beneficial in analyzing data and accumulating findings, then realizing how data sets are just numbers and learning how I can turn those into the writing that I love to do, was really powerful,” she said.

Emma Hazeltine, a junior at Marquette, said, “I got involved after I learned this project and the DNC wasn’t just for political science students.” As a Freshman at Marquette, Kaley Gilbert, says developing research skills early in her education is key. “To be able to look at and analyze data and convert it into something meaningful to this election has been the best part of this project,” Gilbert said.

Even through the COVID-19 pandemic, the students referenced the collaborative nature of the project as a key takeaway. “Being on this team, staying connected even though we’re all different ages and have different classes, and working with UWM and the NMDSI is really cool,” Hazeltine commented.

 

About the Northwestern Mutual Data Science Institute
The Northwestern Mutual Data Science Institute is an industry and academic partnership between Northwestern Mutual, Marquette and UWM formed to inspire and cultivate passion for data science in the Milwaukee region. Leveraging the strengths of the three institutions, the groundbreaking partnership will contribute $40 million over five years to help build a technology ecosystem and advance southeastern Wisconsin as a national hub for technology, research, business and talent development, while creating an organic pipeline of tech talent in the area.

To learn more about the Northwestern Mutual Data Science Institute, visit nmdsi.org.