Pinki Mondal - University of Delaware Master of Data Science -

By RJ Braun

The Data Science Institute (DSI) at the University of Delaware sponsored a full-day symposium at the UD STAR Campus in STAR Tower Audion on February 12, 2020. This event was a formal kick-off meeting for the DARWIN High Performance Computing (HPC) resource. In addition to a few invited talks, there was a poster presentation session with posters from faculty and students within the UD campus as well as from external partner institutions.

Two MSDS students are currently supervised by Dr. Pinki Mondal.  Dr. Mondal is an assistant professor in the Department of Geography and Spatial Sciences, a resident faculty member of DSI, and an affiliated faculty member with the MSDS program.  She gave a rousing presentation on land and water use in a variety of locations including India and its neighboring countries, Vietnam, and our own DelMarVa Peninsula.  Among other students, she described the work of Mr. Abdul Qadir and Ms. Vishruta Yawatkar, who are both in the MSDS program.

Ms. Yawatkar has begun getting results on salinity intrusion using airplane aerial imaging with one-meter resolution.    Image processing clearly showed an instance of clear intrusion in Talbot County, Maryland.  Another image of Kent County, Delaware, showed piecemeal intrusion.  Sea level rise could certainly exacerbate such situations.

Mr. Qadir is a coauthor with Dr. Mondal on a 2020 paper in Remote Sensing of Environment that uses USA and European Union satellite data to monitor forest degradation.  They developed maps of different forest and other plant life in India and neighboring countries.  The freely available paper offers strategies for policy makers to most economically and effectively deploy workers on the ground to monitor the forests and to modify policy to most effectively.

Mr. Farid Qamar, who is supervised by Dr. Gregory Dobler of the Joseph R. Biden Jr. School of Public Policy and Administration.  Dr. Dobler has a primary appointment as a assistant professor in the Biden School, is a resident faculty member of the DSI, and is an affiliated faculty member with the MSDS program.  Mr. Qamar’s poster used hyperspectral imaging as a source of data with the potential for rapid and non-destructive assessment plant health, which may be done remotely and in an automated fashion. He examined and compared various machine learning methods performing image segmentation on ground-based hyperspectral imagery in order to distinguish the materials and make use of their spectra for further analysis.

Program director Richard Braun’s graduate student, Ms. Rayanne Luke, presented a poster on parameter identification in tear breakup on the ocular.  Tear breakup parameters can’t be measured directly in vivo, and Ms. Luke’s models automatically minimize the difference between experimental fluorescence images of the tear film and computed dynamics of the fluorescence from mathematical tear film models.  The minimization process identifies the parameters that best describe the experiments.

The MSDS is proud of these students’ contributions to the DARWIN kickoff meeting!