Partner: Co-Insights In partnership with Co-Insights, this DS4CG project explored the #StopAsianHate movement using NLP to analyze topic transitions, identify significant events, and highlight key accounts driving conversations. Unlike prior studies focused on peak activity, this longitudinal analysis examined changes over time, implementing text embedding and clustering models to uncover frequent unigrams, phrases, and example tweets.
Partner: Environment Canada DS4CG 2023. Aerial insectivore populations have been declining in Canada, prompting the need for a model that can predict bird migration patterns. When a roost of birds take off together, they appear as a distinct shape in weather radar data. By adapting algorithms to fit Canadian weather radar data, we created a model that accurately predicts migration patterns. In addition, the model can even detect roosts that were previously missed by EC.
Partner: Red Cross Netherlands DS4CG 2023. The focus of the second summer of our partnership with the Red Cross was to integrate DISCount, UMass Computer Vision Lab's new approach to estimating counts in object detection, into the Red Cross' workflow for counting damaged buildings after a disaster. Our lightweight model helps responders save time and effort in determining the disaster's impact and severity, which ultimately helps the Red Cross deliver aid quickly.
Leverage modern machine learning and AI techniques to rapidly analyze the before and after satellite images of affected areas to assess damage.
When whales are encountered near the surface of the ocean, it’s useful to be able to quickly assess their health for conservation and evaluation purposes. As part of our Data Science for the Common Good program, Master’s students Chhandak Bagchi and Gizem Cicekli developed a tool for identifying a whale from an aerial image.
In partnership with the World Resources Institute, a global research organization that works on issues such as energy efficiency and climate change, we examined vehicle usage in developing countries. There is often a lack of critical data for developing countries on vehicle purchases and usage patterns, which is a significant gap in sustainability research.