Inaugural Run of the Open Source Community Collider a Smashing Success!
Thank you to everyone who joined us for OSSci’s inaugural Open Source Community Collider! We had 20 attendees out of 70 registered, and I feel confident in suggesting we all had a great time zooming around the collision rings.
Participants engaged in a series of 4-minute (increased to 5 minutes after the first round) one-on-one sessions where we met a diverse group of individuals. I personally met folks working in metadiscourse, contributing to tools for fluid simulations on the ISS, people that want to see their projects join NumFOCUS, people releasing AI frameworks and running large hackathons, someone that met their current SO on GitHub, people that very much agree with OSSci's mission to reconnect researchers with engineers, and folks that are beyond thankful for the support NumFOCUS gives their projects... they enjoy getting to work on the software and not the bureaucracy that NumFOCUS abstracts away from them.
Several participants used our prompt questions to frame their rapid discussions. They were:
What brings you to the collider?
Do you contribute to an open source project?
Do you contribute to research?
Have you ever been to a PyData event?
After the collisions, we spent 15 minutes with all the participants taking feedback and discussing how to move forward (everyone thought this was very much worth their time, and wants to participate again).
We also all joined the OSSci slack to stay in touch! Want that invite? Join an OSSci interest group or the collider calls!
What’s Next?
Based on those discussions, we’re excited to increase the energy of these collisions going forward!
For next month we will
Keep the collision time to around 5 minutes instead of the planned 4.
Host two events per month, accounting for different time zones:
September 13th at 7am ET: Register here
September 27th at 2pm ET: Register here
You’re encouraged to join both to meet your community peers from around the world!
Proactively reach out to invite more diverse communities!
Add a shared whiteboard on which we can all doodle and communicate.
Provide access to a shared document to jot down notes, ideas, and contact information during the event.
Get Involved:
Register for September’s Community Colliders!
September 13th at 7am ET: Register here
September 27th at 2pm ET: Register here
We encourage you to invite others from your communities. Bring a friend or two and be the biggest element in the chamber!
NOTE: By registering, you’re consenting to receive one follow-up email with a summary of the past event and details on the following event.
Thank you for helping us make the first OSSci Open Source Community Collider a success. We look forward to seeing you at our future events as we continue to build connections and drive collaboration across the open source and open science communities.
Mark your calendars and register for September 13th and 27th—see you there!
Introducing the Open Source Community Collider!
Note: This is a repeat post as we begin to transition old Medium blog to our website’s updates! The Community Collider takes place tomorrow, August 9th at 1pm ET.
The original Medium post was on July 22nd
Can't make it to all the open source events? Wish that hall track lasted forever? Looking to meet and collaborate with more of the community?
OSSci's Open Source Community Collider is for you! This event is designed to foster connections and spark collaboration within the NumFOCUS and broader open source and open science communities.
Event Structure:
Stage 1: Kick off with a 5-minute overview in the main chamber.
Stage 2: Dive into the collision rings for multiple 4-minute one-on-one calls, switching partners after each session.
Stage 3: Gather back in the main room for a group discussion, sharing insights from our conversations to guide future events.
Open to All:
Everyone in the community is welcome! We’ll provide several prompt questions to guide conversations, but feel free to talk about anything you'd like. The 4 minutes are yours to use as you and your companion see fit.
Future Events:
This is a pilot run! If folks enjoy this use of time, we'll look into organizing these collider sessions monthly to continue building connections and collaboration.
Code of Conduct:
We will be operating under the NumFOCUS Code of Conduct to ensure a respectful and inclusive environment for all participants.
Don’t miss this opportunity to connect, share, and collaborate with fellow community members.
Who knows who you'll bump in to.
When? August 9th from Noon to 1pm ET
Join the 60+ already registered: https://numfocus-org.zoom.us/meeting/register/tZYpf-utrjojHNHR2XRRyBtOrb_X0BKLSQzi#/registration
Mapping UVM’s Contributions to Open Source at the First PyData Vermont
It was an absolute pleasure to be part of the first PyData Vermont on July 29th and 30th. We were joined by Ellie DeSota, co-founder of the pending 501c3 and silver sponsor of the event: The Science Coordination Infrastructure and Operating Systems Collaborative (SciOS).
We took the opportunity to do some live mapping of the Vermont technology and open source ecosystem, and we look forward to expanding it in the coming months!
I helped lead a BoF on MOSS, and SciOS coordinated an hour and a half workshop on needs and outputs related to the map. The BoF gave us two full pages of suggestions I can’t wait to explore, while we left the SciOS workshop with an early sketch of a tangible technology-based use-case for the MOSS database: a security and funding risk assessment dashboard. More on this in the future!
We also took the opportunity to showcase some work we’ve been doing regarding mapping institutional contributions to open source. Find those images at the end of this update. There will be a future post on how exact we’re doing this in the coming weeks, after we experiment with a few more institutions.
Some lessons learned from our first mapping of an academic subsystem in open source:
There is not a lot of continuity across an institution with regards to repos, tags, keywords, practices, etc. This means we will have to personally tailor each mapping for each institution. We will also have to continue iterating, and integrate user verification of relationships soon!
Showing the result of that lack of continuity can leave an impact on an institution (and lead, perhaps, to changes within the institution). For example, toggling the map to view UVM associated repos with a requirements file drastically reduces the number… from over 100 to about a dozen. A citation.cff query reduces the number of projects to 0. Perhaps faculty will begin to encourage their students to include their relevant files in their repos moving forward.
A core hypothesis of MOSS is that demonstrating the value of certain practices by elevating projects, organizations, people, etc. that follow them will encourage best practices to emerge from the bottom up instead of having them imposed from the top down, or suggested without practical reason. This first test-case seems to support this hypothesis.
There is a lot of excitement how MOSS and the database on which it is built can impact institutions.
Projects that have come out of UVM and the domains in which they are used.
Projects that have come out of UVM and the domains and fields in which they are used.
People at UVM that contribute to open source software, and the domains in which that software is used.
The relationships for institutions we are mapping so far. SDG stands for the UN’s 17 Sustainable Development Goals.
The Map of Open Source Science, now in Neo4j. Now in 3d.
Explore the results of our scikit-learn mapping, and see the advancements of MOSS over the past two weeks.
First, a quick slideshow of what happens when we connect the recently added 20,000+ nodes related to scikit-learn to the wider ecosystem.
Scikit-learn on the right, the rest of the NumFOCUS ecosystem on the left.
You can see some connecting nodes that contribute to both the scikit-learn ecosystem and other aspects of the NumFOCUS ecosystem.
Here is one such connection. The edge exiting the right of the screen is going to scikit-learn.
MOSS Updates
Words do not do justice to the amazing developments made on MOSS over the past 2 weeks. So here are some screenshots. None of this would be possible without Mark Eyer’s contributions.
While we work to build out our Neo4j database and subsequent visualizations, we continue to push Kumu to its limits. This is the complete rendering of the interactive map so far: 50,000+ nodes, tens and tens of thousands of relationships.
This is a similar rendering of MOSS, but in the Neo4j browser.
We transitioned from the Neo4j browser via a docker instance to the Neo4j Desktop application and Bloom visualization tool.
The beginnings of MOSS in 3d. The red system is scikit-learn. The orange system is the Bioconductor dependency graph. The green systems are other projects and their contributors.
The first 3d rendering of the NumFOCUS ecosystem.
The above images all show about 50,000 nodes. We are currently processing hundreds of thousands of additional nodes related to three specific systems:
Project → papers that cite the project → authors of the papers → institutions to which the authors are affiliated — just like we did with scikit-learn
Academic institutions → repos associated with the institution
Academic institutions → people associated with the institution
Here’s a preview.
Well over 50,000 projects, papers, authors, and institutions.
200,000 nodes.
We can’t wait to explore the emerging connections of these systems.
We also expect the need to reexamine our data structure in the near future. What a perfect time for you to get involved!
jon@numfocus.org