Jupyter Hackathon Series in Hawaii
PyHawaii hosted a three-part Jupyter Hackathon as part of a workshop grant, generously offered by Bloomberg. PyHawaii, a user group in Hawaii, led the three-part event series composed of a:
- Git/GitHub hands-on workshop (4-hrs — September)
- Jupyter Lab, Notebook and more hands-on tutorial (4-hrs — October)
- Jupyter Hackathon/open source sprint (2-days — November)
In this post, Project Jupyter had announced the availability of Bloomberg funding to host small-scale community events. PyHawaii wrote up a grant application for three events and was approved. The original grant application was for an event with ~25 people. Since the venue was donated by the University of Hawaii, PyHawaii basically needed funds to cover food and incidentals. PyHawaii also included in the application a request for funding to have a Jupyter Core Developer attend the two-day hackathon in November. Ultimately, Project Jupyter was able to send two Core Developers: Dr. Brian Granger and Dr. Ian Rose to the third event.
Based on the strong responses to the initial Meetup postings, PyHawaii sought additional funding via their relationships with Booz Allen Hamilton to provide food for more attendees. PyHawaii advertised via Twitter, Facebook, meetup.com and relied upon their relationships with professors, college students, technical professionals and members of the community to get the word out. PyHawaii was on Bytemarks Cafe, their local tech broadcast on Hawaii Public Radio.
How it all played out…
Each of the events, described below, played in slightly different ways.
Git/GitHub Hands-on Workshop
PyHawaii had a number of mentors and organizers (Jeff, Joe, Jim, John and Chalmer) who walked 35 attendees through PyHawaii’s “How to contribute to open source” workshop. This workshop is used to prep folks for open source sprints and has been run at conferences around the world (twice at Pycon (2018, 2017), PyOhio (2016), DjangoCon (2018) and more).
The material is designed to maximize hands-on use of git and GitHub and enable attendees to practice the most frequently used commands and features by contributing to a repo full of poems (this allows them to focus on just the mechanics of git/GitHub and not have to simultaneously digest all the nuances of an unknown code base.)
Jupyter Hands-on Tutorial
For PyHawaii’s Jupyter Tutorial, we used a 4 hour lesson created and led by Chalmer Lowe, with support from PyHawaii mentors and organizers to introduce 39 attendees to Jupyter Lab, Notebooks, widgets and more. Like their git/GitHub tutorial, this PyHawaii lesson is also open-source and freely available.
Jupyter Hackathon
Over the course of a weekend 25 attendees came together to work on adding features, fixing bugs and improving documentation in Jupyter Lab, Ipywidgets and other repositories. Jupyter core developers (as well as mentors from PyHawaii) spent both days working with attendees to:
- understand the codebase for the various projects
- troubleshoot the process of setting up their dev environments
- identify suitable issues to work on
- craft suitable pull requests
In addition to the hackathon, a meet-and-greet event was hosted for core developers, professors and students at the University of Hawaii.
Outcomes and Lessons Learned
These events achieved success on multiple levels. Dozens of attendees had a chance to experience the fun (and frustration), the excitement (and exasperation) of contributing to open source, and they got to do it in a friendly environment with mentors on-hand to help guide them past the initial hurdles. These new skills and experiences have the potential to shape career opportunities and strengthen bonds in the local developer community.
The attendees had multiple opportunities to network and grow their relationships. Hawaii has seen tremendous growth in their development community over the past few years and events like this help to speed that growth. The attendees had the opportunity to network on an international scale through discussions about possible fixes to issues with core developers in person and via the web. Based on feedback, a number of attendees initially had no idea of the breadth of capabilities in Jupyter. As attendees apply this experience to their work, adoption of Jupyter as a data analysis, data science and interactive, exploratory platform will grow.
So far, PyHawaii has at least 13 pull requests, 10 of which have been accepted!
PyHawaii’s experiences yielded the following lessons learned, should other organizations seek to facilitate similar events:
- a 1:12–15 person ratio of experienced project developer to attendee allowed folks to get quick, hands-on mentoring without over-burdening the developers. Having other mentors (even if they aren’t experienced in the nuances of the project) is a must to support the project developer in staying focused on answering project-specific questions vs more administrative/low-level questions about git, etc)
- host a short event (i.e. the evening OR previous weekend) before the hackathon to help walk folks through existing issues and through setting up their dev environment (build tools, documentation build tools, source code, compiler, etc). This will enable attendees to hit the ground running when the hackathon actually starts
- RSVPs are tricky when you host an event at no cost to the attendee. Depending on the event, we had between 28% and 53% no-shows which greatly impacts food costs, which often has to be purchased beforehand. Regular tracking of attendance at other events can help identify no-show trends and facilitate more effective budget estimates.
All in all, this was a great experience that promises to have a lasting impact on the data science and programming communities in Hawaii.
