Satellite images contain valuable information about city processes such as urban sprawl. This project fit a deep vision model to classify parts of cities into urban classes. The model was then used to measure urbanization over time. By pairing this index with social and economic census data, the end product, a visualization tool for policymakers, aims to identify "strained" counties, areas that are growing in population without the proper resources to support them.

University resources are sometimes hard to locate due to unorganized and clunky websites. This is especially a problem in emergency situations. TXConvergent and UT Student Government partnered to create UText, a mobile app where students interact with a chatbot to access key resources such as the health center, late-night walking services, car breakdowns and campus police. The intiative received funding from UT Student Services.

Improving customer engagement on HomeAway's web and mobile platforms is important for increasing the number of bookings. This project presents a photo intelligence engine that can classify various images in a listing into home interior categories. The model has many uses including organizing listing images, thereby enhancing the browsing experience, and also composing the backbone of a customer service chatbot.