Last updated on April 28th, 2020 at 02:42 pm
While the term “data science” may elicit images of complicated algorithms and artificial intelligence for the layperson, the Northwestern Mutual Data Science Institute is determined to use its research to improve the community.
A partnership of the Milwaukee-based life insurance company, the University of Wisconsin-Milwaukee and Marquette University, the institute launched last fall to prepare students to meet the growing need for data scientists across various industries. The initiative is backed by a $15 million investment from Northwestern Mutual, and $12 million commitments from both schools.
In its first round of research projects, the institute’s team of faculty and student researchers is focusing on three significant societal issues: inequity in a Milwaukee neighborhood, the opioid crisis and 2020 voter sentiment.
“We always wanted to use data science to address social issues; we did not want it to be an esoteric field,” said Purush Papatla, co-director of NMDSI.
The team has partnered with Walnut Way Conservation Corp., a nonprofit in Milwaukee’s Lindsay Heights neighborhood, to analyze nearly 20 years of data on the social, environmental and economic aspects of the neighborhood. It will use those findings to determine how to address barriers to equitable housing and economic development.
“The ultimate goal is to be able to understand what is suppressing home ownership and stability of the community and what can increase it? The bigger goal is to take the learnings from here and see if we can apply those learnings to address the same kinds of issues in other communities in our area and perhaps even nationally,” Papatla said.
Another research project is aimed at understanding the opioid epidemic. Led by faculty members from UWM and Marquette, the research team is using public data from local sources and geospatial analysis to study opioid overdoses in Milwaukee. The goal is to find insights that may be applicable to other communities across the country.
“We felt data science has the power to at least help us make inroads, if not solve (the crisis),” Papatla said.
A third team of researchers is applying a technique known as “social circulation” to multiple data sources, ranging from traditional political polls and debate transcripts to political advertising and social media interaction, with the goal of understanding which issues are on voters’ minds this election season.
“We felt data science could do more than what traditional polling (has done),” Papatla said. “Polling, of course, has been the mainstay to get into the electorate’s mind, but it also, in a way, is limiting in terms of how many times you can go back to people to keep talking to them, getting data back and analyzing it. Between polls, things could change. Opinions could change. New issues could come up. What better way to tap into that than literally the real time data that’s coming from online and social media?”
That real-time data has allowed NMDSI researchers to track, for example, the recent pivot in political discourse from other policy issues to the coronavirus. While civil rights and discrimination, health care and the economy dominated the political discourse on Twitter in February, by the second week of March, users had turned their attention to COVID-19, researchers found.
“With the election project, they are literally scouring the web to see what people are saying,” Papatla said. “They are getting, every day, tens of millions of data points. … In addition to that, the data itself is different from traditional numeric data, in the sense that it could be textual data, people posting on Twitter; visual data, people uploading a picture of something on Instagram; video data, audio data. Data is now very different from the traditional view of what data was.”
While the deluge of data available today has dramatically changed research practices, the challenge is developing the skills to mine meaning from it, Papatla said. But, for companies that are successful at doing so, data science can be used to make more informed decisions, solve problems and create new products.
In a recent panel discussion at the Milwaukee Engineering Research Conference regarding the role of big data in industry, NDMSI executive director Keri McConnell said it now informs all areas of business at Northwestern Mutual.
“In the life insurance industry … the actuaries will say they were the original data scientists, because they were,” she said. “And now when you look at what we’re doing, from underwriting to running our operations, to putting in new systems, creating a rich customer experience with digital, data is in everything we do.”
Papatla said he wants students to apply their skills to fill that need among area companies.
“There aren’t enough people to take advantage of all that data science can offer,” he said. “… As this institute takes root and becomes more and more successful, the region as a whole will become more attractive for employers looking for data science talent. They will come here and set up shop.”