Big Data Analytics & Methods
Big data, with its volume, velocity, variety and veracity (https://www.ibmbigdatahub.com/infographic/four-vs-big-data), provide critical information that can be used to understand health patterns and healthcare utilization, cost, patient outcomes, and consumer behaviors. We use a variety of big data methods to analyze data of different types and from different sources for public health research. Examples are using natural language processing on electronic medical records (EMR) to identify surgical site infections, analyzing consumer reports on social media (e.g. Facebook) to evaluate satisfaction with hospital and nursing home care, and the use of statistical machine learning methods to predict patient outcomes such as 30-day readmission rate and suicide death rate. In addition to the EMR and social media data, the Department can support faculty and student research with more than 15 years of hospital, nursing home, and assisted living claims data as well as large and diverse longitudinal population datasets including the Health and Retirement Study and those within the Multi-Ethnic Study of Atherosclerosis and the Framingham Heart Study.
Program Faculty
Diana Fernandez, MD, MPH, PhD
Our group is exploring the use big data methods to link household supermarket purchases to food composition data bases. Grocery store purchases could be used to monitor longitudinal food and nutrient patterns in the population, to evaluate public health nutrition policies and programs for health promotion and for the prevention of diet-related chronic conditions across the lifespan.
Philip K. Hopke, PhD
I have an extensive background in factor analysis, pattern recognition, and related chemometrics
Yu Liu, PhD, MPH
My research in this area includes the use of social media mining and machine learning to explore HIV-related social media posts mentioned by men who have sex with men to identify predictors of HIV prevention messaging design and HIV incidence in the U.S.
Helena Temkin-Greener, Ph.D., M.S.