OPAL
I worked at the MIT Media Lab in the Human Dynamics group with Dr. Alex Pentland and Dr. Thomas Hardjono on a new security protocol called the Open Algorithm Project (OPAL). The basis of the project was to create a protocol where one could run useful data computations on sensitive data without actually seeing the data. As a result, we maintain privacy while producing a useful data analysis.
I worked specifically on the use cases of this new protocol. I used Twitter and Fitbit as the data providers. Twitter was an interesting use case because it was a live stream of data being provided. I focused solely on the location metadata because I wanted to demonstrate the dangers of Twitter’s public data. I used LDA to determine what topics in certain locations were being talked about. Then, I applied a OPAL back end to show how Twitter could benefit from OPAL. Lastly, I created an algorithm to asses the health of Fitbit user by using their data. AI computations ran on this personal data and outputted a health score while maintaining the data private.
Most importantly, the beautiful people I got to meet at the Media Lab:
