TGIF: Think-Grapple-Innovate-Fridays

June 17, 2020

TGIF Recording

In early 2020, the School of Data Science instituted a new tradition we call TGIF, or “Think-Grapple-Innovate-Fridays.” These recurring events are intended to bring together researchers from a variety of departments and disciplines to explore areas for collaboration. 

In early March, the School of Data Science held a TGIF to discuss data science and the arts. As the COVID-19 pandemic unfolded, a TGIF was held to bring together researchers from the Schools of Data Science, Medicine, Engineering & Applied Sciences to share their perspectives and plans for pivoting their agendas to contribute to finding solutions to the disease’s spread and economic and social impacts.

In response to the murder of George Floyd and the ensuing nationwide protests, the School of Data Science hosted a TGIF on June 12 to bring together faculty from Data Science, Law and Public Policy with members of the community to explore the intersection between data, ethics and police misconduct. 

Several faculty members at the School of Data Science and others from across Grounds shared their research and perspectives.

UVA Wikimedian-in-Residence Lane Rasberry presented data collected by Wikipedia community members on protests related to the death of George Floyd. He shared Wikipedia’s map which pinpoints where protests have occurred throughout the world, when they occurred, and how many people attended. This map can be found here

Charlottesville resident and software engineer Malcolm McLachlan and Detroit-based Frank Romo shared their own efforts to crowdsource and visualize data about protests.

Deborah Hellman, Roy L. and Rosamond Woodruff Morgan Professor of Law, discussed algorithmic bias and gave several examples of algorithmic bias in policing. One example was determining an individual’s likelihood of going to jail. She explained that this is often determined by prior arrests; however, it is crucial to take into account that prior arrests incorporate not only a person’s previous crimes but also policing practices. Hellman pointed out that minority groups are often policed more heavily than other groups, and therefore they are more likely to go to jail. 

Hellman also posed questions about technology and how these tools must be used ethically.

“How much should the past control the future?” Hellman asked. “If AI and big data allow the past, a past with injustice, to control the future more than the past already controls the future, what are we going to do with these tools? How can we be sure to avoid compounding and procedural injustice?”

Luis Felipe Rosado Murillo, Researcher at the School of Data Science and Center for Data Ethics & Justice, teaches the Data Ethics course to MSDS students. He detailed what this course entails. This year they studied mass incarceration. Specifically, the students looked into the technology that estimates a person’s likelihood to commit another crime and how this is used to automate decisions about bail and parole. Murillo explained that this technology often incorporates bias. 

Michael Porter, Associate Professor in the School of Data Science and the School of Engineering and Applied Sciences, shared some publicly available data sets on police interventions, misconduct and shootings and highlighted the limitations and potential uses of these data.

“We need better data on policing,” Porter said. “So much is not reported. What should the School of Data Science do now? Should we be storing data sets in this area? How can we prioritize this going forward?” 

Don Brown, Senior Associate Dean for Research and Quantitative Foundation Distinguished Professor in Data Science and W.S. Calcott Professor of Systems and Information Engineering, shared some of his research on hazard prediction as employed in Afghanistan and proposed this approach as a way of forecasting and preventing incidents of police misconduct.

Sarah Burke, a Charlottesville mitigation specialist and private investigator who specializes in capital murder cases, also joined the call. In the aftermath of the Unite the Right rally in August 2017, Burke was appointed by the Charlottesville City Council to help establish the first Police Civilian Review Board. This CRB worked to develop an oversight model that would review the Charlottesville police. The CRB requested the past seven years of data on stop-and-frisks, arrests, etc. There have been improvements in transparency since then, including public reports on arrests right after they were made, but, Burke asserted, there is a long way to go.

Larry Terry, the Executive Director of the University of Virginia’s Weldon Cooper Center for Public Service, joined the call to challenge the way that universities can often become ivory towers, cut off from the rest of the world, not taking tangible action. He asked how the University of Virginia might use data on policing in a pragmatic way. Terry was recently appointed by Governor Northam to the State Crime Commission.

Raf Alvarado, Director of the Masters of Science in Data Science program, moderated a discussion among panelists and fielded questions from audience members, many of which focused on next steps that the School and University would take to foment collaboration around the issues discussed.

Dean Bourne opened and closed the TGIF event with his own reflections on the actions the School will take to support both Diversity of our field and the application of data science to issues of injustice and racial inequality. These include the creation of staff and postdoctoral positions, the establishment of a scholarship fund, and both capstone and graduate fellowships dedicated to topics related to the intersecting crises we face as a nation. This TGIF is only the first of a series of conversations that will rally resources and expertise across grounds and beyond to support movements seeking justice, equality and inclusion.