Did you know that you can navigate the posts by swiping left and right?
Today we concluded our third year of a virtual summer internship program in data science!
Because of COVID, we moved the internship online in 2020, and we’ve kept it online both due to the ongoing risks of COVID as well as the increased access the online option has brought.
We kept the same technology stack this year as last year, doing almost everything on Discord and our JupyterHub except for faculty lunches/invited presentations, which were on Zoom (since we can count on anyone external having a Zoom account at this point). As before, we split interns into pods of ~3 with a student mentor who had been through the program before.
Our major format change this year was the expansion of the “programming bootcamp” that last year was just the first week (really 3 days since the first day is largely overview). This year, we expanded the bootcamp to 2 weeks. In general, the interns felt the bootcamp was helpful, though we don’t have precise learning measures (and we have a small N) to understand how much they really learned or whether it was better to do the bootcamp over interleaving basic programming with data science, as we’ve done previously. One disadvantage of expanding the bootcamp is that we had to take out a week of learning content on more advanced topics (gradient boosting, regularization/penalization, etc.), though obviously advanced topics without a strong foundation is not a good idea.
Like last year, the interns had 2 weeks for projects sourced from community partners. We were fortunate this year to have very clean datasets donated, so 2 weeks for the projects was fairly reasonable (though an extra week could have been used well). One project focused on nonprofit donations and the other on racial inequities in traffic stops. Interns on both teams were really motivated, and I was very proud of what they accomplished. Here is a COVID-style group photo I hacked up!
The DataWhys Project and internship are supported by the National Science Foundation through Grant 1918751 for to the University of Memphis.