Researchers Leverage Data Science to address Critical Illness and Injury

By | Research

Critical illness and injury is a silent epidemic that impacts more than 5.7 million Americans every year.

It has an enormous societal and economic toll, and in recent years, progress in critical care research has grown more reliant on the ability to gather, store, search and analyze big data.

Two teams at the University of Michigan — the Michigan Institute for Data Science (MIDAS) and the Michigan Center for Integrative Research in Critical Care (MCIRCC) — are partnering to find new and innovative ways to monitor, diagnose and treat critically ill and injured patients.

“More and more people are embracing data science tools and techniques to address important challenges in today’s society, ranging from poverty and mobility to health care,” said H.V. Jagadish, MIDAS director and the Bernard A. Galler Collegiate Professor of Electrical Engineering and Computer Science. “Critical care accounts for nearly 40 percent of hospital costs and patient hospital days, so we’re constantly looking for ways to harness data platforms to accelerate research and solutions in this important field.”

MIDAS, established in 2015 as part of the universitywide Data Science Initiative, aims to advance cross-cutting data science methodology and applications, promote the use of data science to benefit society, build a data science training pipeline and develop partnerships with industry, academia and the community. MCIRCC, established in 2014, brings together integrative teams of scientists, clinicians and engineers to develop and deploy cutting-edge solutions that elevate the care and outcomes of critically ill and injured patients.

MIDAS and MCIRCC have coupled their collaboration with external funding to catalyze several multidisciplinary research projects.

In November 2018, U-M mathematician Harm Derksen and a team of researchers from MIDAS and MCIRCC secured a $1.4 million grant from the National Science Foundation. With federal support, U-M researchers are working to design efficient, numerically stable and computationally feasible algorithms for tensor analysis that could be relevant to a wide range of big data applications, including the treatment of sepsis. This project also will help MIDAS develop new interdisciplinary courses on big data.

“The role of data science in health care is rapidly growing,” said Kayvan Najarian, a professor of computational medicine and bioinformatics who also serves as associate director of MIDAS and MCIRCC. “No longer an exotic or novel approach, it is quickly becoming another tool in the toolbox for researchers and clinicians, a methodology deployed deliberately to serve a defined research need.”

Najarian and Derksen, along with MCIRCC Executive Director Kevin Ward, also partnered to help develop the Analytic for Hemodynamic Instability, which was licensed to Fifth Eye, Inc. in 2017 and has since raised $11.5 million in Series A funding. Using analytics from a single streaming EKG lead, the tool can predict if a patient will deteriorate several hours before normal vital signs signal a problem is occurring.

“Projects like these highlight the importance of using data science to help patients by providing insights on the challenges they face and when to take action to meet them,” said Ward, a professor of emergency medicine who also is a member of the MIDAS executive committee.

U-M SI and MIDAS faculty Ceren Budak among first to study Facebook data

By | Research

Ceren Budak, U-M SI assistant professor and MIDAS researcher, is among one of the first research teams to have access to anonymous data from Facebook. She will be studying social media’s impact on democracy in the United States. The study will look at how sharing behaviors on Facebook are affected by changes Facebook makes to the platform. More information can be found here: https://www.si.umich.edu/news/university-michigan-researcher-among-first-study-facebook-data.

The Effect of Social Interaction on Facilitating Audience Participation in a Live Music Performance

By | Research

This research was supported by funding from the Michigan Institute for Data Science.

Title
The Effect of Social Interaction on Facilitating Audience Participation in a Live Music Performance

Published in
ACM, June 23-26, 2019

Prepublication
https://web.eecs.umich.edu/~wlasecki/pubs/crowd-in-c_CC2019.pdf

Authors
Sang Won Lee, Aaron Willette, Danai Koutra, Walter S. Lasecki

Abstract
Facilitating audience participation in a music performance brings with it challenges in involving non-expert users in large-scale collaboration. A musical piece needs to be created live, over a short period of time, with limited communication channels. To address this challenge, we propose to incorporate social interaction through mobile music instruments that the audience is given to play with, and examine how this feature sustains and affects the audience involvement. We test this idea with an audience participation music system, Crowd in C. We realized a participation-based musical performance with the system and validated our approach by analyzing the interaction traces of the audience at a performance. The result indicates that the audience members were actively engaged throughout the performance, with multiple layers of social interaction available in the system. We also present how the social interactivity among the audience shaped their interaction in the music making process.