Selected Projects

Many people, especially those with chronic health conditions such as iritable bowel syndrome and chronic migraine, track health related variables with the hope of discovering connections between causes and symptoms and enabling more informed choices regarding their health. Gaining actionable insights from this kind of data is known to be difficult because current tools often fail to automatically analyze data in scientifically rigorous, helpful, and actionable ways. Additionally, current tools often fail to account for lapses in tracking, evolving health goals, and the growing burden of tracking more information. This tool uses Bayesian network analysis framework to support individuals in analyzing their personal health data in evolving real life contexts.

Current dementia and Alzheimer’s Disease onset predictions rely on medical imaging tests, which are expensive and draining on already scarce medical resources. We are developing a suggested protocol and machine learning prediction methodology that will use a few simple tests to determine if someone is likely to get a dementia diagnosis in the near-term future. These simple tests can be administered in 15 minutes by a physician or assistant at yearly physicals for aging patients. Our system enables prediction of future diagnosis as well as prediction explanations, enabling patients, healthcare providers, and loved ones to take potentially preventative actions, allocate future medical resources effectively, and get a jump start on planning for the future.

A customizable surgical anesthesia monitor using D3 based on needs identified by interviews with doctors. Displays real time waveform vitals data, along with past vitals data trends. Allows users to explore past and present waveforms side by side, customize which waveforms and trends they want to see, and automatically calculates vitals statistics not available in current displays.


. Opportunities for Bayesian Network Learning in Personal Informatics Tools. Virtual Workshop on Artificial Intelligence for HCI: A Modern Approach at CHI’20, 2020.

PDF Project