Third annual MIDAS research symposium emphasizes an interdisciplinary approach to data analysis
Big data, data science and analytics were among the main topics discussed at the third annual Michigan Institute for Data Science daylong research symposium Wednesday at Rackham Auditorium and the Michigan League.
MIDAS is a product of the University of Michigan’s Data Science Initiative, an effort launched in 2015 to invest $100 million in data science research and education across campus. Three years later, the symposium offered a chance for MIDAS to welcome leaders in data science research and showcase the work of University students and faculty.
The daylong event featured a lineup of speakers whose research in data science represented fields ranging from psychology to biostatistics.
Alfred Hero, co-director of MIDAS and professor of electrical engineering and computer science, spoke about how this diverse set of speakers aligns with the theme of the symposium, “A Data-Driven World: Potentials and Pitfalls.”
“This is the first time we have a theme that is really focused on a very broad basis of the future impact of data enabled scientific inquiry, data enabled commercial and lending practices and data in society in general,” Hero said.
Hero explained how the application of big data –– from transportation to the health sciences –– presents a similar challenge.
“Data is being collected about you and other people that is being used for purposes that nobody can predict,” he said.
These challenges were the focal point of the event as the five speakers, including the keynote speaker, data scientist Cathy O’Neil, grappled with the outlook for big data research applications and the potential difficulties that may arise.
Among the speakers was James Pennebaker, a University of Texas at Austin psychology professor, who discussed the integration of data analysis into his social psychology research of words. Pennebaker’s research applies data analysis of pronouns, articles and prepositions to environments such as trauma writing therapy, AOL chat rooms, college admissions essays and email correspondences.
Pennebaker’s experience blending data science with social science research allowed him to gain insight into the dynamic between the two fields and how the future of data science is being shaped by collaboration.
“I think the best social psychology is now frankly being led by Facebook, Google and Microsoft,” Pennebaker said. “What is beginning to happen now is we are getting these two groups of people who are now at this point where we are able to start talking to each other and take advantage of generations of really solid social science that has not been informed by giant data.”
Information graduate student Kristen McGarry was in attendance and echoed Pennebaker’s sentiment specifically with the collaboration between the research and corporate spheres. McGarry noted how the corporate sponsors of the symposium such as TD Ameritrade and Mercury Fund demonstrate the corporate interest in data science.
“Data is growing so much, you kind of need both people to join together to continue progressing,” McGarry said.
Rackham student Jeff Lockhart, assistant organizer of the symposium, expressed how this interdisciplinary future of data science is a central mission of MIDAS and the symposium.
“MIDAS has twin goals of getting researchers together in the same room from different disciplines and … it funds a lot of projects,” Lockhart said.
Lockhart described how MIDAS has been successful in bringing people together and funding projects in its short existence.
A poster showcase at the event featuring more than 50 research projects across the University relating to data science emphasized the scope of the symposium.
The MIDAS Initiatives Panel featuring leaders within MIDAS’s research initiatives continued the symposium’s interdisciplinary focus.
Carol Flannagan, the principal investigator of MIDAS’s Transportation Challenge Project, spoke during the panel about how it has helped her research on transportation safety to include perspectives from the social sciences, statistics and engineering.
“One of the very cool things about what MIDAS is supporting and encouraging is these very large multidisciplinary teams,” Flannagan said. “I am working with people from all over campus, from different schools and colleges and people I have not worked with previously.”
MIDAS members are excited for the future of the institute as the symposium came to a close at the institute’s new space in Weiser Hall.
While Hero accounted for the weather being a key reason attendance was lower than expected, he also expressed excitement for the global audience of more than 4,000 viewers the symposium attracted via livestream. A sign, he believes, of the growing number of people interested in the work of MIDAS and the future of data science.