For a long time now, I’ve been thinking about three big challenges in open science:

• Coding is hard enough by any measure – coding for sharing & reuse is even more demanding. Given that our traditional education system isn’t yet imparting these skills to scientists & researchers, and given that it takes sustained practice over a long time to integrate these skills into our research, how can we help build those skills at scale?
• Many students and early career researchers feel intensely isolated and unsupported in their efforts to learn to code, leading to fear of embarrassment before their colleagues, struggles with imposter syndrome, and uncertainty on how or even if to proceed with their research careers.
• The production of open source software in support of open science is not enough on its own; we also need to lower the barriers to discoverability and collaboration so that those projects actually get reused, as was done at the NCEAS Codefest last year – but we need to do it at scale and at home, without requiring expensive trips to conferences.

At some level, these are all the same problem: they are all endemic to a fragmented community. Taken all together, the scientific community has a huge amount of programming knowledge; but it’s split up across individuals that rarely have the opportunity to share that knowledge. Crippling self doubt often arises not from genuine inadequacy, but a loss of perspective that comes from working in isolation where it becomes possible to imagine that we are the worst of all our peers. And as we saw at the NCEAS event, the so-called discoverability problem evaporates very quickly with even a small group of people pooling their experience.

The skills & knowledge we need are there in pieces – we have to find a way to assemble them in a way that elevates us all. The Mozilla Science Lab thinks we can do this via a loose federation of Study Groups.

Our Powers Combined

I started thinking about Study Groups last Autumn, after a conversation with Rachel Sanders (PyLadies San Fransisco); Sanders described regular small PyLadies meetups where learners would support each other as they explored a tutorial, project or idea, where emphasis was on communal, participatory learning, lecturing and leadership roles took a distant back seat, and learning occurred over the long term. By blending these ideas with something like a journal club familiar to many academics, I think we can build Study Groups that powerfully address the questions I started with. I’d like Study Groups to do a few things:

• Promote learning via a network effect of skill sharing. By highlighting the authentic, practice-driven use of code, tools and packages led by the researchers who actually use them in the wild, we create an exchange of skills that scales, grows richer and tracks real scientific practice the more people participate.
• Create and normalize a custom of discussing code as a research object. Scientists and researchers need forums where the focus is on code and the methodologies surrounding it, in order to create space for the conversations that lead to discovering new tools and improving personal practice.
• Acknowledge the ongoing process of learning to code by putting that learning process out in the open & making it shared among colleagues, in order to dispel the misconception that these skills are intuitive, obvious, or in any way inherent.

In practice, these things can be achieved by getting together in an open meetup anywhere from once a month to once a week, where individuals can lead follow-along demos, have a co-working space to explore and experiment together, and everyone feels comfortable asking the group for ideas and help.

Predecessors & Beta Tests

A number of powerful examples of similar groups predate this project, and I had the good fortune to learn from them over the past several months. Noam Ross leads the Davis R User’s Group, a tremendously successful R meetup that has generated a wealth of teaching content on R over the past few years; Ross also organized a recent Ask Us Anything panel on the Mozilla Science Forum, and invited the leads from a number of different similar programs to sit in and share their stories and experiences. I met Rob Johnson and others behind Data Science Hobart while I was in Australia recently; DaSH is doing an amazing job of pulling in speakers and demo leaders from an eclectic range of disciplines and interests, to great effect. And I’ve recently had the privilege of sitting in on lessons from the UBC Earth & Ocean Science coding workout group, which informed my thinking around community-led demos on tools as they are actually used, such as Kathi Unglert‘s work on awk and Nancy Soontiens‘s basemap demo.

From these examples and others, I and a team of people at UBC began discussing what a Study Group could look like. For the first few weeks, we met over beers at a university pub, in the Hacky Hour tradition started by our colleagues in Melbourne at the Research Bazaar. Enthusiasm was high – people were very keen to have a place to come and learn about coding in the lab, and find out what that would look like. Soon, with the help of many but particularly with the energetic leadership and community organizing of Amy Lee, we had booked our first event; Andrew MacDonald led a packed (and about 2/3 female) room through introductory R, and within 24 hours attendees had stepped up to volunteer to lead half a dozen further sessions on more advanced topics in R from their research.

There was no shortage of enthusiasm at UBC for the opportunities a Study Group presented, and I see no reason why UBC should be a unique case; the Mozilla Science Lab is prepared to help support and iterate on similar efforts where you are. All that’s required to start a Study Group at your home institution, is your leadership.

Woot hoo a great turnout for our first #rstats group. Cc: @billdoesphysics @MozillaScience @polesasunder pic.twitter.com/K9oNwiXEVm

— Amy Lee (@minisciencegirl) April 15, 2015

Your Turn

In order to support you as you start your own Study Group, the Mozilla Science Lab has a collection of tools for you:

• We’ve built a template website using GitHub Pages that you can fork and remix for your own use. Not only is the website served automagically from GitHub, but we took a page from Nodeschool.io, and set things up to direct conversation & event listings to your issue tracker, thus adding a free message board & mailing list. Check out the Vancouver R Study Group‘s use of the page; setup instructions are in the README, as well as on YouTube – and as always, feel free to open an issue or contact us at sciencelab@mozillafoundation.org if something isn’t working for you.
• We’ve written a first draft of the Study Group Handbook, that pulls in lessons learned from other groups and guides newcomers through the process of setting up their own, including a step-by-step guide for your first few events, lesson resources, and more. This is a work in progress, and it’ll only get better as more people try it out and send us feedback!
• We have begun to collect lesson plans & resources delivered in similar meetups for reuse community-wide. If you’d like to maintain your own lessons, send us a link and we’ll point to your work from our Study Group Handbook; if you’d rather we do the maintenance for you, send a pull request to our collection and we’ll make sure your work helps elevate the entire community.
• Finally, get on the map! Whether you start a Study Group with our tools, or you’re in one running on its own, send us a link and a location and we’ll add you to the map of Study Groups Worldwide, so others in your community can find your meetup, and we can all see the global community that is emerging around working together.

We’re very much looking forward to working with you to help you spool up your own Study Group, and learn from your experiences on how to make this program what the research community needs it to be; we hope you’ll join us.

Study Groups, Hacky Hours & Open Science Meetups