Hackathon

The RobustiPy hackathon is to be held from 12:30pm on the 27th of June at Said Business School. Directions on how to get there can be found here. The hackathon will in particular take place in Seminar Room A: There should be a member of our team waiting in the lobby helping to direct people to Seminar Room A from between 12:15-13:00.

A finalised schedule of events for the day is as follows:

12:30 - Tea and coffee on arrival with an optional drop-in session for installing RobustiPy.
13:00 - An introduction to the day.
13:05 - Introducing RobustiPy.
13:30 - RobustiPy in action (we provide some empirical and simulated examples).
14:15 - Pizzas!
14:45 - Bring Your Own Dataset or work as a team to test your data and (try to break!) RobustiPy!
15:45 - Take a break and make a slide or two to present back to the group.
16:15 - Present your findings, and prizes for best application!
17:00 - How can we improve RobustiPy?
17:45 - Roundtable discussion on the future of multiversal analysis.
18:15 - Social events (subject to interest) and further discussion!

We encourage participants to try to install RobustiPy using the instructions found on the GitHub page here. If you are unable to install RobustiPy for any reason, kindly email us in advance, or arrive in the ~30 minute pre-arrive session so that we can help with any installation issues (see above).

Participants will get the most use out of the session if they bring a dataset with them. We suggest a simple cross-sectional tabular file with either a binary or a continuous dependent variable, with one or more specific ‘x’ variables of interest. Generally, unless you want and have access to run this on a High Performance Computer, the dataset which you bring should have <10 control variables.

We would also kindly request registrants to consider reading the three papers listed on the ‘Inspiration’ section as described here:

Simonsohn, U., Simmons, J. P., & Nelson, L. D. (2020). Specification curve analysis. Nature Human Behaviour, 4(11), 1208-1214.
Young, C., & Holsteen, K. (2017). Model uncertainty and robustness: A computational framework for multimodel analysis. Sociological Methods & Research, 46(1), 3-40.
Gelman, A., & Loken, E. (2013). The garden of forking paths: Why multiple comparisons can be a problem, even when there is no “fishing expedition” or “p-hacking” and the research hypothesis was posited ahead of time. Department of Statistics, Columbia University, 348(1-17), 3.

These three papers – along with the results of our discussions on how to improve RobustiPy – will form the basis of our final session on the future of multiversal analysis.

Finally, you should be able to find some slides relating to the day here. This single slide deck contains administrative information, acts as the ‘handout’ for all sessions, and contains all other general information that you should reasonably need on the day (including links to resources, slides, etc).

We look forward to seeing you there!!