top of page

Projects

Find all of my publications at Google Scholar

Evaluating Fair Machine Learning Toolkits

Fairness researchers have developed toolkits to help machine learning (ML) practitioners incorporate fairness into their work, but there's a gap between these tools and practitioners' needs. We conducted a mixed-method user study to develop a rubric for evaluating and enhancing future Fair ML toolkits. Find the paper at the ACM Digital Library. 

fairness.png

Exploring Music Through Genre Visualization

We prototyped a web-based graph visualization called TastePaths that helps expert music listeners, such as editors and curators, explore new artists within nearby genres. Our user study suggested that expressive ways to explore can help music listeners better shape and provide feedback on their recommendations. Find out more at the Spotify Research blog. 

tastepaths_system_v4.png

Giving Voice to Silent Data

Streaming services personalize recommendations by leveraging data about people's listening habits, which we term as 'silent data.' We ran a series of exploratory participatory design workshops to find out how music listeners would want to interact with this data in conversational interactions. Find out more at the Spotify Research blog. 

music-data_sm.gif

Eliciting Recommendation Feedback by Voice

We designed and deployed an online experiment on a crowdsourcing site to explore the potential design space of using a voice assistant to elicit feedback to improve listener recommendations. Our data suggested that people prefer a collaborative or teaching relationship with their voice assistant. Find out more at the Spotify Research blog. 

assistant-capabilities.jpeg
bottom of page