I spent the day with design firm IDEO and a large number of scientists and stats people in San Francisco, finishing up a two-day meeting to design a program in Data Science supported by the Moore and Sloan Foundations. It was odd to interact with designers in a scientific context. Their ways of operating and thinking and communicating are very alien for scientists. However, they did a good job—via a well-developed and thought-out process—of bringing out important questions, ideas, and common themes for institution-wide and multi-institution interdisciplinary research in Data Science. Now we all head back to our home institutions and write proposals, while the Foundations and IDEO prepare some public documents about principles of design pertaining to Data Science initiatives. Odd!
Some of the most interesting ideas that came up involve going meta (applying data science to understand data science, using profiling to assess programmers and scientists and not just programs) and mash-ups (Foursquare plus TripAdvisor plus github plus Python packages). I hope we implement some of these ideas and they have legs.