I actually just wrapped up my course today. The coursed is called Materials Informatics. We had the final presentations this past Monday for the newest and last NSF-IGERT Ph.D. program at Georgia Tech. This program FLAMEL is geared towards designing new and improved materials using data science approaches. In the class, students were paired into teams and given brand new research projects with internal and external faculty collaborators. They had the term to do research and give a presentation summing up their contributions.
The problem sets spanned applications like nanomaterials, high-performance titanium, fiber composites. The datasets that the students were operating on were substantially large, as is most materials image data. Also, the students ranged from entry-level to senior graduate students; some students spend a lot of time at a lab bench in excel; some contribute to open source projects. The diversity was incredible.
The 9 Project Pages.
Time was split between technical applications of data science and being responsible contributors to science through the web. This focus:
- taught a minimalist syntax and software
- advised efficient data communication methods
- promoted frequent blogging of research results
- enabled asynchroneous collaboration
- tracked provenance
- never talked about
The students produced conference quality presentations on difficlut research projects within the span of the term. The collaborations collaborated through code, data, blogs, and in class discussion/teaching.
A lot of our course information is now persistent and discoverable. I ran some analytics using the Github API from a single repository, I'll get those codes up soon. There is a summary of the information aggregated during the course.