Last week, we hosted our first global Mozilla Science Lab sprint. Over 22 cities were represented in the sprint (18 official sites with a number of other individuals joining us virtually) for two jam-packed days of round-the-clock sprinting on tools, lesson materials and resources for open science. It was a fantastic way of uniting the open research community around the world to collaborate on projects that further science on the web.
The sprint started in New Zealand, with sites in Auckland and Wellington kicking us off on Tuesday, July 22 (their time). From there, we passed the #mozsprint baton to community members in Melbourne, Australia, then to Europe as they began to wake up. We continued to pass the work on to the east coast of the United States as Europe broke for lunch, working until Europe signed off for the evening just as San Francisco and Vancouver got rolling. Then we started the cycle once more with our colleagues in New Zealand.
Timezones included, over the course of the two days, we clocked an astounding 52 hours of sprinting, linked between sites by chat clients, video screens and github issues. An astounding amount of work was accomplished during the sprint, and we wanted to take a moment to highlight some of what was worked on. You can find out more here in the project etherpads, as well as a roundup of activity here. We’ll be featuring a series of guest posts over the next few days to shed further light on the sprint, and hope you’ll join us in keeping up the momentum started last week on these projects. Still plenty of ways you can get involved.
On to some of the highlights:
Lessons and learning materials:
- Training materials for data on the web (“Data Carpentry”) saw a flurry of activity. Check out the project etherpad for more, from lessons on Excel, social science examples, Pandas to infrastructure for the Data Carpentry project. Many thanks to Tracy Teal, Christie Bahlai, Aleksansdra Pawlik, Hilmar Lapp, Karen Cranston, Amy Nurnberger and others for their leadership and help.
- Bioinformatics capstone examples for Software Carpentry, led by Rob Beagrie, Liz I-S, Stephen Turner and Scott Ritchie. The team made two lessons, aimed at aspiring bioinformaticians and aim to pull together what they would have learned in a novice R bootcamp or a novice python bootcamp.
- Medical imaging capstone examples, led by Matthew Dimmock. The aim of the capstone is to teach people how to select regions of data in medical images (CT, PET) for fitting and processing. The pull request (see bottom of page for latest modifications) can be found here. You can access (or fork) the IPython Notebook Matthew’s been working in here.
- Oceanography capstone examples, led by Damien Irving and Nicolas Fauchereau. For more, check out this pull request.
- Digital Blacksmith course notes, led by Paul Mignone, to explore an training course approach for 3D printing and CAD. You can access the project and notes here. Additional content will be added in coming weeks. Currently focusing on the #Blender tutorials (06_AdvancedCAD.md)
- Hard-hitting guide to using Excel for science by Aleksandra Pawlik, Christie Bahlai, and many others.
- Improved Spreadsheets lesson (also noted above at ‘Hard-hitting guide to…’), thanks to Alexander Duryee and many others.
Tools for teaching:
- Tool to ease some of the administrative burden in organizing Software Carpentry bootcamps, led by Atul Varma, Neil Chue Hong, Sam Penrose, Amanda Harlin, Ernest Walzel, Matt Davis, Phillip Sipe and others. Together, the team added an event list page, the ability to filter events by site, worked on documentation, added a csv bulk uploader tool and modified the homepage search UI. This all builds on initial work done by Atul – fantastic to see where it’s come since that hack first started.
- Browsercast – Caleb Creed worked really hard on understanding the scope and scaling for Browsercast – an IPython Notebook plugin which facilitates lecture recording and playback to assess the project.
- Peer Instruction Teaching Tool, an effort to build a web application that implements the peer instruction teaching technique using the WebRTC technology and PeerJS project. This work was led by Piotr Banaszkiewicz, Rémi Emonet, Raniere Silva, and Leszek Tarkowski. You can see a full list of tickets closed as part of the sprint here.
- Introducing Scientists to Testing and Code Review, a “mentored open source contributathon” experience led by C. Titus Brown and Michael Crusoe at Michigan State University. A #mozsprint tour de force by many, debugging our dev install/test process, with the 13th pull request merged right at 17:00 local time marks the official end of the sprint. Nine contributors, including five working remotely – all new to the project. Many thanks to Heather L. Wiencko, Jeramia Ory, Ivan Gonzalez, Justin Lippi, Josh Herr and more. Also, have a look at what the sprint looked like from a project perspective.
- Import Lesson, led by Remi Monet and Raniere Silva tried to break Software Carpentry lessons into small pieces to be reused by anyone. Thanks to W. Trevor King.
Tools for better science:
- ContentMine toolset, led by Ross Mounce, Peter Murray-Rust, Richard Smith-Unna, Noam Ross, Stephanie Unna, and Robin Berjon. ContentMine is a project to liberate facts en masse from scientific literature via automated text mining. The tools provide a framework for researchers to extract material such as DNA/Protein sequences, identifier systems (e.g. Genbank, ProteinDB, DOIs, equipment, etc.) from journal articles.
- Scholarly markdown, on using Markdown to author scientific documents. Edward Abraham and Christopher Knox went through the process of taking a published manuscript and preparing it in markdown, to test out scholarly markdown. Martin Fenner also began work on a website for scholmd. The conclusion was that scholarly markdown is missing intra-document references, and requires better support for templating, tables, and some other more minor fixes.
- NeuroStars – Paulinešek Kafičko implemented ORCID OAuth2 support for Django. Roman Valls implemented Magnet support for bittorrent (scientific) data sharing.
- Modular File Renderer: Rendering, Exporting, and Sharing Scientific File Types, led by the team at the Center for Open Science. Over the course of the two days, they tackled a number of bugs, nearly completed ipython module, completed pdf module, planned geospatial module
- Open Notebook Science Network, working to create a network of open science notebooks, especially to close gaps between WordPress-based notebookers and researchers who are more technical or computational e.g. who prefer GitHub based collaboration. The team sprinting on this included Brian Glanz, Egon Willighagen, Colin Closek, Lisa Crosson, Rolf Pankratz, Parker Moore, Ben Balter and more.
- JotGit, a science-focused git-backed real time collaborative editor project, had various contributions, from some fixed meteor package bugs, progress on file handling, auto-save and file uploads. Many thanks to Theodor Vararu, Desislav Hristov, Bernhard Konrad and John Lees-Miller for their help! You can access the main repository here.
- Reproducible research guidelines, led by Stephanie Wkystra, Ben Marwick, Bruno P. Kinoshita and Daisie Huang, including contributing a new code-writing guidelines section to the rOpensci reproducibility guide.
Many thanks again to all of the site hosts, project leads and participants. We hope to continue building momentum as well as ties within the open research community, furthering the work started last week. Want to help us carry this work forward? Get in touch. Also, a reminder that the call for proposals is open for this year’s Mozilla Festival. Want to build on one of these projects at the Festival? Get that application in now.