Students will complete a semester-long research project in 3-4 person teams. Projects must satisy three requirements:

• address a real, design problem in a particular domain
• use a data-driven approach to model the problem
• evaluate the effectiveness of the approach

Beyond these requirements, students are enouraged to be creative, and should think broadly about domains and datasets .

Possible domains include graphic design, online advertising, print media, visualization, 3D modeling, architecture, interior decoration, fashion, software programming, education, protein design, cooking, etc. Really, any domain of human endeavor has design problems!

Similarly, there are many different datasets to work with. More and more design is done digitally and stored on the cloud; therefore, many interesting repositories of design can be scraped from the Web. Alternatively, crowdsourcing is an effective way to collect data that isn't readily available. Also, there are probably a host of private datasets that people would be willing to share for the right project, so ask around!

The goal here is for students to pick and execute on a research project that could be submitted as a UIST Poster in early July.

A list of your final project group members and your project's abstract. The abstract should clearly define the research problem/question, your approach to modeling the problem, and plan for evaluation. We will provide feedback to help you better scope the problem.

Short 10-minute presentations on your project, which will include real preliminary results.

The final write-up should follow the CHI Late-Breaking Works' Extended Abstracts format. The write-up should be six pages long (not including references).