Application submitted to https://science.mozilla.org/blog/2018-mini-grant-rfp by Tim Head tim@wildtreetech.com in February 2018.
Our objective is to increase the number of people publishing educational, journalistic, and scientific materials that enable readers to read, edit, re-run and re-use that content. We believe Binder is the tool to achieve this.
We plan to create the material for a one-day workshop giving an introduction to the Binder project and the public mybinder.org service, and using it at three test events in Europe. These first events will allow the material to be trialed and improved, so creating a lasting curriculum and allowing others to run similar workshops. The workshops will serve educators, journalists and scientists who want to learn to use Binder to improve the accessibility, shareability and reusability of their work. The three workshops will be held in locations where we have local support: Cambridge, UK, Berlin, Germany and Lausanne, Switzerland. We will promote the events and materials within the wider python and R communities as well as research and open-data networks we have access to.
We will produce a tested schedule and content for a one day workshop on Binder. We will extend existing Binder training material and documentation to include content focused on users and newcomers to the community. We will create building blocks for sessions from which to assemble a workshop. The development of the material, schedule and planning of the events will happen publicly on GitHub. The community creating the materials will be recruited from experienced Binder contributors and members of the workshop target audience who are trail-blazers in their respective fields. This activity will provide them with ideas, a network and material to run their own events. All materials produced will be open source.
The outcome of this project will be an increased number educators, journalists and scientists creating Binder ready repositories that will allow their students, readers and collaborators to start reading, editing, re-running, and re-using their content. By lowering the barriers to experimenting with published analyses, we will promote more open innovation, reproducibility of analyses and efficiency in being able to re-use existing methods. Workshop participants will leave knowing how to create a Binder ready repository for their project. People enthusiastic about Binder will meet like minded individuals, and so growing the Binder community.
Train educators, journalists and scientists who want to learn to use Binder to improve the shareability and reusability of their work.
Re-using existing work that involves the software of other scientists, educators or journalists is currently hard work, and attempting to do so is a daunting task even for technically skilled individuals. Preparing an easy to try out project equally requires a large amount of effort from the authors. For example, today when a scientist sees a computational technique in a journal article that they would like to use for their own project they have to: 1. discover the source code 2. try to locally recreate the computational environment, including all dependencies 3. contact the original authors with questions when step 2 fails 4. try to run the software 5. re-contact the authors because something does not work. In reality these steps will be repeated several times until one can re-run the original code, if the individual is persistant enough to even get this far. This turns the potentially half-hour job of evaluating how suitable an existing piece of work is for re-use into a multi-day endeavour. The success rate of re-using suitable software is therefore low. As a result tools are often reimplemented from scratch or researchers continue to use pre-existing and outdated tools because they are already setup. This slows down the spread of new ideas, wastes time on setting up dependencies, and discourages or frustrates authors as they feel that if they attempt to share their work they will need to become a help desk to get users running their software.
The Binder project drastically lowers the bar to sharing and re-using software. For the user to try out someone else's work requires only clicking a single link. For the author to prepare a binder-ready project is much easier than having to support many different platforms and for many projects involves little additional work. However Binder itself is a tool that requires some investment to master. The majority of researchers, journalists and educators have not yet discovered the Binder project or have heard it mentioned but do not know how to get started. Software engineering is not their discipline and so mybinder.org is seen as yet another tool they need to learn. This means trying out Binder is put off until some myhtical point in the future when they have spare time. Ironically this means Binder suffers from exactly the problem it solves for other projects.
List key project activities (what will you do), outputs (what will be produced through your activities, e.g. products, publications, number of workshops held and people trained) and outcomes (impact of your project on your beneficiaries during the project period).
We will create the material for a one-day workshop giving an introduction to the Binder project and the public mybinder.org service, and use it at three test events in Europe.
The first part of the project is the creation of the training material, establishing a community of contributors, and planning the in-person workshops. To create the material we will build a community of contributors following the methods of the Mozilla Open Leadership Training. The community will be recruited from experienced Binder contributors and members of the workshop target audience who are trail-blazers in their respective fields.
The second part of the project will focus on planning and running three one day workshops and incorporating the experience of the workshops into the material. These first events will allow the material to be trialled and improved, so creating a lasting curriculum and allowing others to run similar workshops. We will promote the events and materials within the wider Python and R communities as well as research and open-data networks we have access to. The three workshops will be held in locations where we have local support: Cambridge, UK, Berlin, Germany and Lausanne, Switzerland. Each workshop will be limited to 20 participants so that they can be very hands-on and interactive.
The output of this project will be a tested schedule and content for a one day workshop on using Binder. As the material will be available under an open license from day one we hope people unrelated to this grant will reuse it to hold their on workshops or tutorials. Over the course of the three workshops we will have directly trained 60 people in the use of Binder.
The outcome of this project will be an increased number of educators, journalists and scientists creating Binder ready repositories that will allow their students, readers and collaborators to start reading, editing, re-running, and re-using their content. Workshop participants will leave knowing how to create a Binder ready repository for their project. Contributors to the material will gain experience working in an open project and explaining complex technical ideas to novices. By participating in the creation of workshop materials, they will gain ideas, a network of like minded and material to run their own events. People enthusiastic about Binder will meet like minded individuals, so growing the Binder community. We will extend existing Binder training material and documentation to include content focused on users and newcomers to the community. By lowering the barriers to experimenting with published analyses, we will promote more open innovation, reproducibility of analyses and efficiency in being able to re-use existing methods.
Project outcome: an increased number of educators, journalists and scientists trained in Binder
Indicator: Attendance figures from the workshops
Project outcome: Participants from the three workshops create a binder-ready repository that is used in the time after the workshop.
Indicator: We will ask participants for their GitHub username and monitor the publicly available usage information of mybinder.org for launches of Binder ready repositories associated to these usernames. We will contact workshop participants at the end of the project asking for stories about successes and failures in using Binder since the workshop. The mybinder.org service collects data on the number of launched binders, which is made publicly available. We will monitor this data stream and try to attribute a change in usage patterns to the workshops and online training material this project created.
Project outcome: Benefits for contributors
Indicator: By using GitHub as a platform to create the material we will collect information about the number of contributors to the workshop material, how often the material is accessed, forked, and starred.
Project outcome: Growing the Binder community
Indicator: material added to the Binder training and documentation pages by workshop participants
The primary beneficiaries of this project will be the educators, journalists and scientists who attend the workshops in order to learn to use Binder to improve the accessibility, shareability and reusability of their work. The project benefits those who already have a need for a tool like Binder, but don't yet know it exists. For example journalists who publish their R notebooks alongside their articles, a teacher of a course that relies on software tools that need setting up, an organisation that promotes the use of open data, or a scientist who publishes the data and source-code for their latest paper. Each of these authors already publishes their work but re-use and interaction with this material is limited because readers have to install software tools to use it. This limits the reach of their work.
We will engange them by offering a potential solution to their problem of readers having to setup software environments. We will explicitly contact "local heros" who strive for this behaviour and are near to one of the three workshop locations, asking for recommendations for who should know about the event. We will also advertise at local events and on notice boards. Furthermore we will make us of our online presence to advertise the event.
The secondary beneficiaries of this project are the individuals who will form the starting point for the community creating and maintaining the workshop materials. They already have experience using Binder, but not in contributing to an open source project. We will reach them by looking at the public data stream from mybinder.org to find relevant repositories, as well as users reporting themselves as examples to the Binder team, of which Tim is a member. We will offer them the opportunity to showcase their work by including it in the workshop material and ask them for "lessons learnt".
The project has many potential indirect beneficiaries within the fields of education, journalism and science, in those who will benefit from future workshops, better training materials and the extended Binder community.
The creation of the workshop material will take place globally as it is open to anyone who wants to join. Like the current Binder community which is distributed across several continents, we hope to draw on talent from around the Internet. We will hold three in-person workshops in Europe; one in the UK, one in Germany and one in Switzerland.
List any risks or challenges that may affect the overall success of your project, and note how Mozilla and/or others can help you to overcome these challenges.
A primary risk is that not enough people contribute to the workshop material. A potential way to mitigate this risk is to change the focus of the workshops towards creating the material. Mozilla can help with promoting the effort to create training material and the workshops within the Mozilla Science Lab community.
A secondary risk is that no one will register for the workshops. Mozilla, Project Jupyter and others can help by advertising the event within the relevant networks. As well as sharing know-how on reaching the target audience.
A further risk is we do not manage to attract people with the skills required to create appealing material for target audience. Mozilla can help by connecting the proponents within the network of Mozillians where the skills required are present.
A risk is that the workshop participants and material creators will disperse without forming a community. Mozilla can help with the experience of building the Open Science Lab and Open Leadership Training cohort communities.
A final risk is that the project budget is too small. The grant currently financing part of the project Binder development could contribute the missing money. This idea has been discussed informally in the Binder team.
A key partner is the existing Binder community, whose role it is to provide answers to the technical aspects of Binder and ensure a smooth operation of the service on which the workshop depends. They are skilled in the technical aspects of Binder, and maintain the public mybinder.org service.
Local contacts are also key project partners. For the three workshops we have partners at the locations: Naomi Penfold (eLife) in Cambridge, Peter Grabitz (Charite) in Berlin and Luc Henry (EPFL) in Lausanne. They will help with the mechanics of organising a local event (venue, food) as well as promoting the event locally to journalists, scientists and educators.
Mozilla works in the open. How will you document and share your project progress with the community?
All material will be developed in the open under a CC-0 or CC-BY license and hosted on a public GitHub repository. Issues will be tracked on GitHub. Discussions will take place in a public chatroom. If any voice/video calls meetings take place minutes will be created using an etherpad which will be posted to the related issue on the GitHub repository. We will share the progress of the project with the community by publishing blog posts on the official Project Jupyter blog. Workshops will be open to anyone who wants to attend (within the attendance limits).
Are other organizations working toward outcomes similar to those described in this proposal? If yes, explain how your work complements that of others or fills a key gap.
The team behind the Binder project are creating documentation for their project and giving talks/tutorials at conferences. To date the material is focussed on how to contribute to the project itself and setting up your own instances of the Binder project. One section of the documentation (https://mybinder.readthedocs.io/en/latest/using.html) details how to create a Binder ready repository. It is written with technical readers in mind. Creating introductory tutorial material is part of the long term roadmap of the Binder project. Tim is part of the team that maintains Binder. Current contributors to Binder are focussed on developing the software and operating the public mybinder.org service. To cover additional areas at the same level of quality the Binder project recognises that it needs to grow the community of contributors and encourage projects that are "outside" of Binder that add to the community. Therefore this effort has the full support of the Binder project.
The question is confusing, since I apply for this grant as an individual and not an organisation. Personally I am committed to open source software and increasing levels of data analysis and statistical literacy.
The project contributes to a healthy Internet by encouraging consumers of "data driven products" like newspaper articles or scientific publications to look behind the scenes of those articles. By making it easier for authors to publish the material that lead to an article and easier for a reader to explore that material we will increase the number of people who do so. This will increase the number of people who at first inspect the material behind an article, then feel confident enough to make small changes and eventually create and pubish their own work. By increasing the number of individuals who ask questions about research and reporting we will increase the number of people who understand how valuable the data they share online is. Thereby allowing them to make more informed decisions about sharing it or not.
Is this a new project or a continuation? If new, please describe your qualifications to initiate the activity. If continued, please describe your accomplishments to date. Feel free to include links to articles and documents online that highlight your recent work.
This is a new project. There are currently no efforts specifically targeting educators, scientists and journalists to help them get started using Binder. However the proposed project is related to the larger Binder project which has recently been rebooted. The mybinder.org service served 10,000 users from 120 different countries over the 30 day period ending January 25th 2018. The launch of Binder was announced on https://elifesciences.org/labs/8653a61d/introducing-binder-2-0-share-your-interactive-research-environment and https://blog.jupyter.org/binder-2-0-a-tech-guide-2017-fd40515a3a84 and it was recently featured in https://www.nature.com/articles/d41586-018-01322-9.
Tim is a central member of the Binder team. He has organised workshops and training events before, including the on-going LHCb starter kit. This was designed to train new members of a scientific collaboration in in-house software, and continues after three years, despite Tim and other original instigators moving on. See https://lhcb.github.io/starterkit/. Tim also leads the scikit-optimize open source project and community. This is a three year old project with 35 contributors and 670 "stars" on GitHub. See: https://github.com/scikit-optimize/scikit-optimize.
Tim has invested his own resources in the Binder project and will continue to do so. Growing the community of users is critically important to finding a sustainable funding model for project Binder and mybinder.org. Therefore both Tim and the Binder project have a vested interest in maintaining the material and community created during this project.
A healthy Python data ecosystem is in the interest of Tim's employer who relies on it for their business. They allow Tim to spend work time on open-source projects.
Have you previously partnered with Mozilla? Have you previously received a grant from Mozilla? If so, please list title, amount, and purpose.
Tim has been a mentee in the first Open Leadership Training cohort and a mentor in rounds since. No other partnerships or grants.