WHERE SILICON HITS THE SIDEWALK: RESPONSIBLE DATA SCIENCE AT PITT
Michael Colaresi
Creating and analyzing data—data science—should not be isolated in a computational silo. Anticipating a world of data applications across the University community, Pitt’s Responsible Data Science Initiative is bringing data science to every field across Pitt, including those that have not typically used computation but are now generating and influenced by data on an unprecedented scale.
Data science is about people, explains Michael Colaresi, associate vice provost for data science and leader of the Responsible Data Science Initiative.
“We don’t simply focus on computation in and of itself. We focus on the decisions that the models and the algorithm help to guide,” he says. “The aim is decision making that is data driven but human steered.”
“The promise of responsible data science is the product between qualitative knowledge and computation and data. And like any multiplication, if either one of those factors is zero, the answer you get is zero.”
Beyond areas like health sciences and engineering, Pitt’s data initiative brings in business, social sciences, policy and the humanities. Applied data science in social sciences is familiar to Colaresi, whose background is political science. He also is the William S. Dietrich II Professor of Political Science, served on Pitt’s data science task force and co-founded a major in computational and social science in the School of Computing and Information.
“Pitt can be uniquely impactful on the applications of data science. We are where silicon hits the sidewalk,” says Colaresi. “We apply broad expertise in foundational data science and computation, biostatistics and artificial intelligence to solve societal-level problems from addiction to sustainability to equity and access.”
Even further, Pitt is connected to organizations in banking and finance, retail, operations, government and the nonprofit sector.
“Sometimes our image of data and computation is glowing blue bits of abstract ones and zeros. Pitt’s perspective allows us to see past this illusion. Our view focuses on how data science can help people and organizations gain agency over computation,” says Colaresi. “It’s wonderful to speed up an algorithm, but we value that speed because of what it unlocks practically.”
Pitt’s Responsible Data Science Initiative is named to intentionally highlight “responsibility” as the core of data science.
“Data and models are important pieces of almost all decision-making puzzles in our current age, but we cannot forget that people are not only the edge and corner pieces but the puzzle makers and players without which there are no solutions,” Colaresi says. “The amazing ability to compute and measure means we need people at the table to help us fit together trust and oversight, surveillance and privacy, as well as convenience with value.”
Pitt is well placed to help resolve these conflicts. Colaresi’s team won’t stop trying to innovate algorithms, but they think about collecting data responsibly and protecting privacy in applied settings. Responsible data science means being thoughtful and intentional about the context in which data could be used as opposed to assuming a one-sized-fits-all solution designed without industry or community input.
“The promise of responsible data science is the product between qualitative knowledge, and computation and data,” says Colaresi. “And like any multiplication, if either one of those factors is zero, the answer you get is zero.”