Do Hackathon Projects Change the World?
Reviewed by Greg Wilson / 2021-10-08
The short answer to the question posed by McIntosh2021's title is "probably not". They analyzed the GitHub repositories of almost twelve thousand hackathon projects from 2018–19 and found that:
…approximately 85% of commits were made within the first month, and approximately 77% of the total commits occurred within the first week. Only 7% of projects had any activity 6 months after the event ended. Evaluated projects had an average of only 3.097 distinct commit dates…
That said, this paper doesn't look at the long-term impact of hackathons on the participants themselves. In my experience, many people participate in hackathons to build new social connections and to learn how to create, review, and merge changes. Those impacts don't show up directly in the hackathons' own repositories; pointers to studies would be very welcome.
McIntosh2021 Lukas McIntosh and Caroline D. Hardin: "Do Hackathon Projects Change the World? An Empirical Analysis of GitHub Repositories". Proc. Technical Symposium on Computer Science Education (SIGCSE), 2021, 10.1145/3408877.3432435.
Hackathons, the increasingly popular collaborative technology challenge events, are praised for producing modern solutions to real world problems. They have, however, recently been criticized for positing that serious real world problems can be solved in 24-48 hours of undergraduate coding. Projects created at hackathons are typically demos or proof-of-concepts, and little is known about the fate of them after the hackathon ends. Do they receive continued development in preparation for real world use and maintenance as part of actually being used, or are they abandoned? Since participants often use GitHub (Microsoft's popular version control system), it is possible to check. This quantitative, empirical study uses a series of Python scripts to complete a robust analysis of development patterns for all 11,889 of the U.S. based 2018-2019 Major League Hacking (MLH) affiliated hackathon projects which had GitHub repositories. Of these projects, approximately 85% of commits were made within the first month, and approximately 77% of the total commits occurred within the first week. Only 7% of projects had any activity 6 months after the event ended. Evaluated projects had an average of only 3.097 distinct commit dates, and the average of commits divided by the length of the development period was only 0.1. This indicates that few projects receive the post-event attention expected of an actively developed project. Finally, this study offers a dialogue of possible ways to reformat hackathons to help increase the average longevity of the development period for projects.