The Changelog – Episode #335

Enabling open code for science at NumFOCUS

featuring Gina Helfrich

Guests

All Episodes

We’re talking with Gina Helfrich the Communications Director for NumFOCUS about their story and history, the impact of open code on science, the difference between sponsored and affiliated projects, corporate backing, the back story of their education and events program PyData, and the struggles of storytelling and fundraising.

Featuring

Sponsors

LinodeOur cloud server of choice. Deploy a fast, efficient, native SSD cloud server for only $5/month. Get 4 months free using the code changelog2018. Start your server - head to linode.com/changelog

Clubhouse – The first project management platform for software development that brings everyone on every team together to build better products. Get an extra two months free - head to clubhouse.io/changelog

Raygun – Unblock your biggest app performance bottlenecks with Raygun APM. Smarter application performance monitoring (APM) that lets you understand and take action on software issues affecting your customers.

FastlyOur bandwidth partner. Fastly powers fast, secure, and scalable digital experiences. Move beyond your content delivery network to their powerful edge cloud platform. Learn more at fastly.com.

Notes & Links

Edit on GitHub

Transcript

Edit on GitHub

Let’s start out with your role in particular. We’re talking about NumFOCUS, but it’s a large organization, a lot of layers to it; “Open code for science” seems pretty deep… Where can we begin to understand your role, and then the origin story of you? Where can we begin here with this conversation around NumFOCUS?

It’s a little paradoxical in that NumFOCUS is a very large organization, and yet there are only about half a dozen staff at present.

Each of us has our official job title(s), but we all do a lot. Oficially, I am the communications director and the program manager for diversity and inclusion for NumFOCUS. What that means in practice is that I help to manage a lot of conversations, and community engagement, marketing, donor engagement, I support the projects with things that they need, and lots of helping people talk to one another on the communication side… And then doing that in a way that feels inclusive and brings in a diversity of people, communities and perspectives on the diversity and inclusion side.

There’s one thing that’s reminded us of diversity and inclusion in the world of science, right? That’s the language of all. It doesn’t really matter if it’s English, or Chinese, or Spanish, or any language you speak - science is a language we can all meet around.

I thought that was music, Adam. Music is the universal language.

Well, science rooted in math, right?

Yeah, mathematics for sure.

I’m being a little bit facetious… Gina, I’m just curious how you came to be at NumFOCUS, and how much of NumFOCUS’s history you have either experienced or been told, since it goes back a ways? Can you tell us about your origin story with NumFOCUS? And even just give us a little bit of the background on the organization itself… By the way, a 501(c)(3) non-profit in the United States. So when we say “official business, big deal, big organization”, that’s some red tape to jump through, so you all have been doing it well… Can you tell us about your history and the organization’s history, if you have it?

[00:04:10.25] I joined NumFOCUS in spring-summer of 2015, and at the time I was essentially employee number two, and came in to help Leah Silen, who was our founding executive director, and still is our executive director. Leah was with the organization since it began in 2012, so I essentially came on about three years in, to help grow and scale.

The way that I got there was just some stars aligning… I am not a scientist, nor a programmer… [laughter]

You’re a doctor.

Yes, I started out in Academia, I have a Ph.D. in philosophy and women’s studies. I did a focus in ethics and social philosophy. After my degree, I spent some time in higher education administration, I directed the Women’s Center at Harvard, and then for a variety of mostly personal reasons I thought I would just make a really big change in my life.

I moved to Austin, I got a job at a technology startup, I was with them for a year, and then as [unintelligible 00:05:34.19] to happen with startups, there was a round of lay-offs and I was among them… And from that point, I sort of had a conversation, heard about something was going on, and then was introduced to Leah. She thought that my Academic experience and my somewhat recently-acquired tech industry experience would serve really well with the needs of NumFOCUS as an organization and as a community, because it very much has a foot in both worlds, if you will.

I think it’s interesting, your history into it, because it didn’t sound like you were apologising that you weren’t a programmer or a scientist, but then Jerod said that you’re a doctor… I think that’s interesting, how Leah saw that in you, because I can see how you’d be good at being in communications and bringing people together. Getting people to that need to talk is often a very hard job, and sometimes it takes a keen eye or a certain personality, and philosophy obviously is a great background to have when connecting people… Because that’s sort of the crux of relationship, is that.

Maybe. I definitely think that my philosophical background and training has prepared me well to interact with computer scientists, because there’s (I think) a very methodical, structured approach to certain kind of problem-solving that I can relate to, even if the application is somewhat different. I have really enjoyed the work that I’ve gotten to do at NumFOCUS as a translator of sorts…

I think one of the really amazing and really cool things about the work that I’m focused supports is that it is highly diverse in terms of scientific discipline, and other applications; not everything that we do actually falls neatly within science… But the different types of work that’s going on is just so interesting and cool, and even though someone who is working in physics might be using some of the exact same tools as someone who’s working in fluid dynamics or astronomy, they still are coming from really different backgrounds and points of view… So I get to have a lot of fun helping people to understand one another better.

[00:08:21.23] We know NumFOCUS is an umbrella for a bunch of member projects, and that you provide services and different things - we’ll get into the details… But just to name a few projects for the listeners out there who maybe heard of NumFOCUS but aren’t sure what’s all involved, I’ll just list off a few that I was well familiar with… Project Jupyter for the Jupyter notebooks, Matplotlib, Julia (the programming language), pandas, rOpenSci… There’s tons of them.

I always in my head associated NumFOCUS with Python, and of course you’ve got NumPy, PyMC3… I can search here by Python and there’s a bunch of Python things. Did it start with Python scientific stuff? Because now you have R, you’ve got Julia, there’s Javascript to this stuff involved… Give us a little bit of where NumFOCUS began, where its roots are, and then we can learn how it grew into what it is today.

As I mentioned, I joined in 2015, so I was not there at the time, but the history that I’ve gathered of what happened, in broad strokes, is that a number of people who were really interested and invested in questions of sustainability for open source projects got together and said “We need to find a way to address what we see as a really acute challenge that’s coming down the pipe… And one of the ways that we wanna do this is by founding a non-profit organization to support these projects.”

Those people who were involved, most of whom ended up being on the inaugural board of directors, happened to be very deeply connected to a variety of Python data tools. The author of NumPy, for example, Travis Oliphant.

So I think that, as I understand it, a lot of the early projects that came on were sort of through those personal connections of the people who were talking to one another, and so that meant that it was a lot of Python projects, but I believe that we did get some other language projects in still relatively early, in the first few years. I think we had rOpenSci for a good long while, but I couldn’t tell you off the top of my head what year they joined, for example. And Julia as well, honestly.

Very cool. So feature-wise, the things that these libraries do - I’m just trying to give like a profile, and you can do it better than I can… If you had to profile a project that would fit inside of NumFOCUS’ umbrella - maybe it’s in there now, maybe it’s not… It’s science-related - that seems like a broad umbrella, but is that a requirement, that it relates to science somehow?

It ends up being something of a family resemblance at this point. I think that the boards looks to approve projects that have some relationship to the general ecosystem of projects that we already have on board. That is not really by language, or a particular feature application, but really more to do with “Is this an open source tool that is being developed by a robust community and has some important application in a variety of scientific fields, usually?”

[00:12:10.26] We have some interesting projects that are a little bit of a different type than just your typical software library or package. The Journal of Open Source Software, for example, is a different kind of project. In fact, somewhat in line, I suppose, with our open side that I mentioned. So we have these projects that provide peer review of software, and other projects, for example QuantEcon, that do online archives and reference resources (they have kind of a library of notebooks), so it is all really focused on various software-based, largely scientific application open source projects.

Gotcha. The biggest aspect of support you have is the fiscal support. You also have this list of affiliated projects, which if I’m reading correctly are very similar, at least in scope or intention, and they fit a certain criteria, but they aren’t fiscally supported by NumFOCUS. I’m seeing a few of these here that I think we’ve covered. First of all, we did have a conversation, Adam - or I did, at least - back at the original Sustain Conference, with Karthik Ram about rOpenSci; that was very interesting, cool stuff, and you can go back and listen to that one.

And then Spack - Spack is on the list. Our friend Todd Gamblin, who was on Request for Commits, is at Affiliated Projects. Why the distinction here between the sponsored projects and the affiliated projects?

This gets into the technical reasons for being of NumFOCUS as a non-profit. One of the really important roles that became clear early on in the existence of NumFOCUS was that open source projects needed administrative support to help provide and give best practice insight into a lot of what has to go into making a sustainable project. NumFOCUS exists to provide that, so that the people who are running the projects can focus on the things that they love and do best.

In order to accomplish that, we started this fiscal sponsorship program, which is not what it sounds like to any normal person… [laughs] Fiscal sponsorship sounds like NumFOCUS has a big bag of money that it gives out to people, but in fact that is not the case.

That’s not how it works…?

There are many non-profit organizations in the U.S. that provide fiscal sponsorship. What it means is that the project enters into a legal contract with NumFOCUS, that effectively makes NumFOCUS the legal entity of the project. To get into the weeds of it a little bit, there are actually two different flavors of fiscal sponsorship. One is the more common, which is a comprehensive fiscal sponsorship, which is really what I’ve just described. We’re the legal entity, for all intents and purposes NumPy is NumFOCUS and NumFOCUS is NumPy, and then there are fiscally sponsored projects which have what’s called a grant or a grantee sponsorship… Which means that NumFOCUS becomes responsible for the use of grant funds within a project, in alignment with its non-profit purposes.

[00:16:32.28] So it’s almost more like oversight to ensure that funds are being utilized in accordance with the non-profit mission… But the main difference between the sponsored projects and the affiliated projects is, as you said, they look really similar on the face, but affiliated projects don’t have any contractual relationship with NumFOCUS, whereas sponsored projects - NumFOCUS is the legal entity for those [unintelligible 00:17:04.25].

What’s the benefit of being an affiliated partner/project?

We have a few benefits… One, most people just wanna be associated with the organization and have access to the network of folks. Being a convening ground I think is one of the really valuable services that NumFOCUS ends up providing to the community and to the maintainers and leaders of these projects.

There are also a few more concrete benefits… NumFOCUS does have a small development grants program, where we make little pots of money available to both our sponsored and affiliated projects, in the form of small grants, to accomplish all kinds of things that their project might need… Whether that’s paying a developer to build out a new feature, or paying for some consulting on how to make their docs more usable, or paying to fly all of the lead developers to a conference where they can get together and have a sprint… There’s all kinds of ways that the funds can be used, but only NumFOCUS sponsored and affiliated projects are eligible for those… So that’s another benefit.

So essentially access to your network, and events and different things that are available to – because NumFOCUS is probably bringing a lot of people to the open code and science yard, so to speak, so with that you wanna be affiliated, but you may have your own legal entity, or in that grant or grantee model you mentioned.

Right, right. Scikit-learn might be a good example here - they’re an affiliated project, but they have their whole own foundation now.

One thing I’m curious about just from an end donator perspective, and maybe again, we’re just getting into the weeds, but hey, let’s just hang out in the weeds for a minute…

When I support NumFOCUS, so I clicked on the Donate button on the website, and there’s “Make a Donation”, 292 supporters (recently, or something), and I can donate - very typical, “Choose your frequency, Choose your donation level”, this is supporting NumFOCUS the umbrella? Does this money trickle down? Is this like the ongoing operations? Because I did find “Check out our other campaigns”, which seem to be more project-focused. How does the actual Donate work?

[00:19:47.01] The primary Donate button on NumFOCUS, if you go to NumFOCUS.org and click the red Donate button, sends you to a form where you can contribute to what is essentially our general fund. It could be used in a variety of ways. The general fund money is what goes to support those small development grants that I’ve just mentioned.

Gotcha.

The general fund might also be used to pay for things like printer paper… [laughs]

It’s general.

It’s one of these things like “How do you chase down a specific dollar?”, it’s very difficult to do. But that being said, the specific Donate buttons on any of the pages that talk about the projects give you the ability to restrict your money to that project. If you went to NumFOCUS.org/project/shogun, you can make a donation to the Shogun Machine Learning Toolbox, specifically to be restricted for their use as they see fit.

Break

[00:20:57.10]

Gina, NumFOCUS has a lot of different arms, or tentacles, or things that it’s doing, and different programs - I guess that’s a fair enough thing to call them… And we talked about the fiscal sponsorship program; you also have a sustainability program… I would lump those into similar things, but help us tease them apart and understand what you’re trying to do with the sustainability program.

You’re exactly right, there’s definitely a close relationship there. The sustainability program is really the effort that NumFOCUS is spearheading to surface and address sustainability issues among our projects specifically. Clearly, a lot of what we learn is gonna probably have resonance for non-NumFOCUS projects, but the purpose of the program is really somewhat focused internally, if that makes sense.

For the past couple of years we have been getting the leadership of all of our sponsored projects together for a summit, to basically have these conversations, understand where things are at, what the needs are, allow folks to do some peer-to-peer mentoring, and understanding how they can learn from one another, and generally just trying to keep building a long-lasting scaffold for the lifetime of these projects.

[00:23:53.06] So you’ve been doing this for a couple of years now… Do you have any key findings or themes that keep popping up, that everybody is dealing with or doesn’t know how to deal with or fails at?

Well, unsurprisingly, we’re always talking about money… [laughter]

I thought they were related…

It’s interesting the different flavors of that conversation, and it depends a lot on the specific project and how it’s situated and structured. Some projects are really in need of money, and they have a very clear idea of what they would do with it. Some projects are wary of taking money, even though they have clear ideas of what they could do with it, but they’re concerned that it would introduce a bunch of friction and open some cans of worms that they don’t really wanna deal with necessarily… And how can some of these larger or more widely used projects receive back some sliver of the productivity and profit that they have enabled other organizations to create because of the type of open source license that they operate under.

If you think, for example, about the amount of money that some place like Google is probably making because they use our tools, it just feels a bit sad that not that much of that money comes back to the projects. While they do support the projects, it’s like, on what ratio, you know?

It’s interesting, because when I think about scientific projects - and this might be an outsider just being naive and wrong, so please correct if I need to be corrected - in the broad scope of open source I feel like scientific tooling-related open source projects would have an easier time convincing their benefactors of their monetary value, like their recompense, than a lot of other ones… Because a lot of firms have big R&D budgets. I guess Academia is a little bit harder in terms of the availability of funding, but there’s grant programs all over the place, and it seems like these things are requirements for success in a lot of the scientific fields, and it seems like we invest in scientific fields at large. Is that just a complete misconception, or are there some that are having some success?

It’s a little bit of a nuanced answer… I feel like we talked about NumPy a lot, but we could take them as an example.

Certainly there are various corporations that have essentially paid people to work on the project, so it’s not like NumPy has never received any kind of support to date… But that being said, they just got their first major grant two years ago, I think, and they have some folks working on it through Berkeley Institute of Data Science (BIDS). There’s a bunch of reasons for that, and some of it is if you’re busy maintaining a project, when do you have time to write a major grant proposal to support the project that you’re maintaining…? But those are some of the types of challenges that we’re trying to untangle through the sustainability program.

Do you offer grant writers support for writing grants?

Not yet. It’s come up, so… We’re talking about money - NumFOCUS needs money so that we can turn around and use it for those kind of things to make more money, right? [laughs]

Right. You’ve gotta have money to make money, that’s what they say out there. You’ve gotta have money to make money.

[00:27:56.05] I was thinking about it too, from your perspective, Jerod, when you were saying “Where does the money come from? Who donates?” and whatnot - I was thinking about the individual projects, the maintainers… Do they end up being independent open source developers? Do they intend to be working at the Googles, the Netflixes, the larger corporations that tend to use this software? Do they end up becoming a sponsor? …like, their time and their maintainership is somewhat sponsored by their jobs. Do you understand what the make-up is, to some degree, of let’s just say maybe your sponsored projects under maintainership, how that tends to be?

Yeah, I don’t think I have anywhere a list of who does what, but I do know off-hand that a number of project maintainers are employed in academic institutions. For example, that is the case for most of the folks who run yt, which was sort of born in astronomy/astrophysics and actually has a grant to expand into a variety of other fields… Whereas pandas maintainers I believe are almost all employed by for-profit corporations. The lead developer for Matplotlib is at a federal research institution, so works for the government, effectively. So it is a mix, but I don’t think that there are that many maintainers represented who are sort of independent. They usually have some full-time position somewhere, and either they are permitted or encouraged to use some of their paid time to work on the project, or sometimes that’s not the case and they’re still doing it in their “spare time”, on a volunteer basis.

Yeah, I was just thinking about that from what it says here in your About page - it says “Our deep expertise in non-profit administration enables project contributors to focus on what they care about most”, and that assumes – I was wanting to know who was maintaining, because the follow-up to that is “development and maintenance of the project itself”, and that seems to be the case for the most part… You want maintainers to be maintaining, but who’s maintaining the maintainers? So that kind of dips into the sustainability [unintelligible 00:30:28.03] money to do these things, grant writers, all these fun ideas…

Yes, exactly.

And then the problem of like “Well, even if they had money, could they use it? Do they really want it?” Jerod mentioned Request for Commits earlier - one of the discoveries in that show was that in a lot of cases money just causes more problems than it actually solves, you know?

Yeah. It’s definitely really complex. We’ve been having this conversation somewhat recently, even with regard to the small grants program. We got some feedback saying “Hey, well one of the reasons my project hasn’t really applied is because I can go to my employer and say that I work on this open source project in my spare time and they don’t have any issues with regard to intellectual property, or things like that, but as soon as I’m getting paid for it, that introduces a conflict… So it’s easier for me to just not get paid for it.” That being said, plenty of projects could happily find ways to spend money if they had it. They could find someone who could come on to work on it full-time.

[00:31:35.25] I think there is a need for this money also to help contribute to solving the challenge of a pipeline for maintainers. One of the things that we’re really talking a lot about lately is that you could have a lot of folks who are making commits and being active, but then there are just a few folks who are really maintaining the project, and reviewing PRs and merging things and making decisions on behalf of their project, and then there’s sort of this valley where it’s quite difficult in fact to have someone move from just submitting a few bug fixes, or even getting to the point of larger pieces of work, and then advancing in their knowledge and sophistication around the project to the point that they can be made a core contributor. It’s almost like succession planning is a really big challenge for the projects, and I suspect that money would help with that a lot.

Well, I think about money sometimes, too - I don’t know if you’ve ever considered this, but this is something that I’ve… I don’t know if this is me coining a term or not Jerod, but I’m gonna say this…“burden offset” — if you are a maintainer of one of these projects, and as you’ve mentioned, if they clearly have a heart for maintaining and advancing and whatnot of the project, and for whatever uses it may have, but they can’t be paid because of just some concerns around that portion of it, that they can offset their burden of maintaining it by being able to use some of these grants or the funds that NumFOCUS can help bring to them, to essentially offset their burden of maintaining it - not so much at the code level, but in adoption; it’s one of the core tenants of open source becoming successful, is adoption, and maybe at that level being a sponsored project. You’ve already attained that, so you’ve already surpassed that need, but further adoption, further education, as you’d done with PyData and whatnot, can become that burden offset.

Yeah, for sure. I think there’s also the case that – [unintelligible 00:33:56.25] projects might get to the point where they realize that they need more people who are part of the core team who are occupied with different types of activities. I think some of our more thriving projects do thrive because they have people specifically whose whole job is to think about these issues, and make sure that newbie contributors feel super-welcome and happy, and facilitate that environment that’s gonna help bring up more people to keep the project afloat… And that might not have anything to do with writing code.

You see that a lot with the Linux Foundation, too… Even at recent – I was at the KubeCon in Seattle in December, and I’m aware of this, but it kind of became more clear to me as I was sitting in this press room with the maintainers of Kubernetes and a couple other projects in the CNCF, and I was just thinking like, aside from the maintainers, there’s so much more orchestration behind the scenes to make this event even possible, to give me and other press affiliates in the room who have a vested interest in the future of software to even be here, and be able to have these people’s attention to ask them questions about maintainership, and code quality, or long-term support of Kubernetes, or just different things that press have questions about - and I hate even calling this press, Jerod, but Sally, that’s [unintelligible 00:35:35.09] We were in the press room, so… You know, whatever.

[laughs]

But there’s so many people behind the scenes that make open source possible that isn’t just maintainers, and it’s unfortunate that some maintainers have the liability of taking money as an issue, but it’s understandable, and so… There we are.

[00:35:55.28] Well, I was just looking at NumFOCUS itself, and Gina, your position, and thinking - you know, we started off with saying this is serious business, this is a sustainable non-profit, this is a big organization, and part of the reason that we feel that way is you all have a communications director. Your skillset is wildly different than other people’s; we talked about your specific skillset, and it’s necessary and helpful, bridging gaps between these people and these people, and all the things that you do, and those are the things that a sustainable open source project that continues to be successful and grow needs over time.

We had a show last week with Kim Crayton, and she pointed out that a lot of the stuff around community and inclusion and codes of conduct etc, that stuff needs to happen very early on, and start with the starters of the project… But the fact of life is a lot of those projects aren’t there, they’re one person working on this thing in their spare time, and they need that help, they need the support structure around them to do things that maybe they either a) don’t have time for, b) don’t have the skills for, or c) don’t have the desire to do by themselves… So there’s lots of that that has to come in and around to sustain something, even once the money problem has been taken care of… Or the money definitely helps with that problem as well, because if you can’t find a volunteer, maybe you can find an employee, for example.

Absolutely. NumFOCUS – once you get to be a certain size or prominence, you’re gonna run into trademark issues, and so NumFOCUS handles that for the sponsored projects. Or you’re running a lot of events, or you have some more complex program for your project community, so now you’ve got invoicing that needs to happen, and accounts payable and receivable, and setting up independent contractors, and things along those lines, and so NumFOCUS handles that. Or you have a complex question about intellectual property, or taxes, NumFOCUS handles that. That’s really where the contribution of the organization is.

One thing that I do wanna call out, just based on what you were describing, is that NumFOCUS has not really designed itself or set itself up to be the kind of place that supports a single developer project, or even getting people from just really starting out to kind of getting going, and the code of conduct support, and that kind of thing. It’s really decided, or maybe just as a virtue of the origins, to be an organization that’s trying to help projects that already have some legs really get to the next level. Our president, Andry Terrel, was just at a meeting the other day where there was a conversation about what it might look like to have an incubator, so something that really is taking in much more early stage projects and helping them to get set up properly, and think about the various aspects that they’re gonna have to have plans for, and kind of structural best practices… But it does not yet exist; it’s just a topic of great interest.

That’s one thing I really appreciate about the CNCF as it relates to Cloud Native - they’ve provided the landscape, but then at the same time they’ve got the incubating projects that eventually have the opportunity to reach adoption or critical mass, and move into a graduated projects scenario and even have more of a bolded/underline name next to the pathway – if you’re familiar with this, they have this concept of a landscape, essentially, of like “Where on the Cloud Native scale does this project fit?”

[00:40:12.02] I think that’s a pretty interesting thing that they have, the incubation process, and then this idea of a graduated process that clearly denotes where a project is, but then at the same time provides the same framework that you said, Gina, that NumFOCUS wants to provide to the budding/early versions of projects, not simply the ones that kind of have hit a more critical mass.

I was just drawing comparisons in my mind as you described that, Gina, to just the startup ecosystem, and you’ve got your angel investors, then you’ve got your seed stage, and then your later stage investors, and it’s like, NumFOCUS is not an angel investor.

It comes in after you’ve reached a certain point. And it’d be cool to have incubators, which would be more akin to angel investments… And that made me think, Adam, of a sweet idea - ready? Open Source Shark Tank.

Uuh, think about it. Just let that marinate.

That could be good.

“Just come pitch your project idea. Maybe somebody will throw a grant your way.”

Break

[00:41:17.13]

Gina, NumFOCUS has a large list of donor names, much individuals supporting NumFOCUS, supporting NumFOCUS projects, we see some business names as well, some supporters, but give us an idea, as far as you know, who all is supporting these projects, who all is using these projects, what does the benefactor (so to speak) look like? Give us some details there.

NumFOCUS gets its revenues in very typical of a non-profit fashion. We do have a corporate sponsorship program; these are the good actors who are getting a lot of benefit out of our tools and have decided to turn around and provide their support in turn.

Do you wanna give any specific shout-outs?

Let’s see… I don’t have them all top of mind, but I could shout out our platinum sponsors. For example our platinum sponsors right now I believe are Bloomberg and Microsoft, and Facebook has just come on, I believe. They’re definitely a sponsor; they might be a platinum sponsor. I can’t remember. [laughs] We’re hiring a development director who will memorize these things. [laughter]

[00:44:13.22] I’m on the sponsors, page - you’ve got Bloomberg, Microsoft, IBM, don’t forget about Capital One, NVIDIA, Netflix, there’s support from the Sloan Foundation, and Gordon and Betty Moore Foundation.

Yeah, I’m realizing that we’re kind of tipping over in the new year and some of those might shuffle around… But yeah, they’re just a bunch of great organizations, corporations who provide support to NumFOCUS to advance our mission.

Our corporate sponsorship program is definitely a really important part of where our money comes from. Another important part of where our money comes from is grants from philanthropic organizations. I don’t think we have yet achieved a recipient of a federal grant, but that’s in process, hopefully. We have gotten a lot of wonderful support from the Sloan and Moore foundations over the years, so we’re definitely very grateful to them.

We do some rolling our own revenues, if you will, through the PyData Conference series. When you got to a PyData event and you pay a registration, in a certain aspect that registration cost is a donation to NumFOCUS, because the net profit of any PyData Conference comes back to NumFOCUS as a means of funding the organization. And in the early days, that was pretty much the only thing funding the organization. The relationship between the PyData program and NumFOCUS is really important, and has that kind of strong historical context.

It’s an interesting perspective too, to think about – I used to be in the non-profit world, but not in the software non-profit world, where a lot of non-profits would find sustainability not just through donors, but being able to offer services to the greater good that would actually be profitable, that would fund themselves. And that’s what PyData is to you.

Right. So that’s definitely a part of it. We wanna make sure that we keep those different funding streams fairly balanced though, so that it never feels like “Oh, this PyData is make or break.” In the old days that was true, but now no longer, thankfully.

As the fiscal sponsor of the projects funding that comes to them through NumFOCUS is also a means of funding the organization. There’s something called an admin fee, which is basically like a small percentage of that money that goes to pay for the overhead of running the fiscal sponsorship program. You can think of it like if a company wanted to come in and give some money to a specific project, like Conda-forge, for example, then NumFOCUS as the legal entity of Conda-forge would accept that money, put it in your marked accounts for Conda-forge, and then a small percentage of those funds would go over into the NumFOCUS general fund, to pay for all the things that keeps us as a non-profit up and running.

[00:47:52.18] Those things that keep us up and running are also the things that provide the benefits of the fiscal sponsorship program. To give a concrete example, those admin fees would go into a NumFOCUS budget to pay the salary of our finance manager, who spends all day every day doing finances for the projects. So it’s all of a piece.

Then last and by no means least, individual donors. You had mentioned a list of names of people who have supported NumFOCUS. We definitely have hundreds of individuals who have very generously made a donation to the organization, mostly because we just get messages that say how much they love the tools.

One of the things I personally feel perplexed and sad about is that I – I’m recently confident that there are millions of people who use NumFOCUS tools all around the world, and know that they are using the tool, so it’s not like a process on their laptop that’s running in the background. They have pip installed Matplotlib, or whatever; they know they’re using the tool, but they have not yet donated back to support that tooling. If you wanna use the Microsoft Office Suite, you have to pay for it; but if you wanna use the scientific Python ecosystem, it’s open source, so you can just find different ways to install it without paying for it… Which is great, and enables a whole lot of things, but what I would hope is that for the people who have the means, they would donate back. You can think of it as like paying for that open source license that you have, just to provide some support to the folks who need it… Because I think the ratio between those few hundred people who donate to NumFOCUS versus the millions of people who use our tools is not really where I would want it to be.

We have a couple of membership programs. You can be a sustaining member of NumFOCUS, where you basically have a recurring donation set up, and it’s up to you how much you want that to be. People do $5/month, people do $500/year, it’s just whatever you want. And we have a little over 200 sustaining members now. That was probably by and large within the past two years that those folks have come on.

Then we also have this idea of a supporting member. If you make a donation of any size to NumFOCUS one time in the calendar year, then you’re a supporting member, and we usually have a couple hundred of those. Those donors are incredibly important, not only for the money that they give, which typically gets put into the small grant funding for the projects, but also just sort of as a representation that the community has the back of the people who are maintaining these projects. So to me, the more that we see the growth of our individual donors, the more that’s gonna be a sign that things are moving in the right direction for NumFOCUS.

Thinking about those long tail of free users who aren’t donating back, do you have any strategies or have you tried anything with regards to moving the needle on awareness, education, really converting those into donors? Because if we throw out a number like five million - if you had five million people using NumFOCUS umbrella projects in the world, well a typical freemium SaaS thing, a conversion rate is I think 1%, Adam; I think better than that is pretty good.

[00:52:02.06] So even if you had a 1% conversion from the people who use it completely open source, free, don’t ever give back, and then you have 1% donor conversion rate - and of course, I’m just assuming 5 million - that’s 50,000 people. So you’re nowhere even in the order of magnitude of that with your individual donors. It seems like if we could move that needle, it could have huge benefits to all these projects. So have you guys tried anything in terms of reaching those folks somehow and turning them into donors?

Yeah, I’ve tried a lot. That’s kind of been my whole job for a couple of years. So the failure, I suppose, is partially mine. [laughs]

Oh, come on…

It’s a hard job. I think if we can all solve this, I think there’ll be a lot of prosperity.

I like the correlation you draw to SaaS though, Jerod, because there’s probably a lot of growth and learning that has happened in SaaS-based software that could be applied to what you’re trying to do, Gina.

Yeah, maybe. I was gonna say, I’m on your show, and in being on your show, I’m trying to do this very thing, which is to reach more users and let them know about NumFOCUS. Most of the time when we do get new donors, they are coming with some version of like “Oh my gosh, I had no idea your organization exists. I use this stuff all the time. You guys are amazing!” So I think it really is mostly an awareness problem, as opposed to “People know, they just don’t wanna donate” - I don’t think that’s it. I think they don’t know that they could donate, and to whom. So just kind of getting the word out I think is a really big one. One of the ways that I’ve been trying to tackle that is to put the tools in the hands of the people, which is to say when someone does donate, in addition to saying “Thanks so much for donating”, we might say like “Hey, why don’t you maybe post to social media and share that you just made this donation, so everyone in your network knows what’s going on?” So trying to kind of get to it that way.

I’m not saying this is the case, but do you think that seeing large brand names as corporate sponsorships could deter individual donors? Because something you said earlier, I think it was in segment one - you were talking about how the Googles of the world, for example, the big four even, are using these tools, and a lot of the money they generate through their services may not be making its way back to… I’m just saying it as sort of a rephrasing of what you said in segment one there, but I’m curious if you thought about it from that perspective… If maybe individual donorship is suppressed because of their assumption that “Well, somebody else much bigger than me, with much deeper pockets than me is benefitting from this software and should be chipping in.”

I had not considered it in those terms. What has come up in the past is that because some of the tools that we support are particularly widely used in data science are perceived as being just so common that it’s sort of incredible for people to think that there’s any issue with the sustainability or support that they’re receiving. It’s like, “Every intro to a data science bootcamp is using these tools, so how could there possibly be an issue?” You’ve got traction, it’s like, “Well…”. Yes, but…

One thing to question too is then how do you drive the awareness of the need for donation? Because the thing you need to attach to that is a “Hey, give us” and “Hey, support us” and then the follow-up to that is “so that…” And what you say after “so that” is the why. As an individual potential donor, that’s the hook that not so much that I need, but it’s what pulls on my heartstrings, is the “so that.” “So that science can prosper”, “So that we can do these things”, and the “so that” is numerous blog posts that come out that share the community’s effort and triumphs around NumFOCUS.

[00:56:28.16] Yeah, yeah. That is something that I think about and sometimes struggle with. When I’m talking to people who have nothing to do with technology, they have no familiarity about my job, they say “Oh, what’s NumFOCUS?”, sometimes I say “Well, it’s an extremely niche non-profit, if you’re not really in the know. And what we do is incredibly important, and the things that are the most exciting are usually multiple jumps down the line from what we do.” So it takes a few steps to connect the work of NumFOCUS to the really exciting things; it’s not such a direct line, and I think that’s one of the challenges that we have around this.

If you’re working for a non-profit that is combatting domestic violence, you can be like “Hey, we’re out there combatting domestic violence”, and everyone’s like “Great! That sounds good!”

[unintelligible 00:57:34.06]

Yeah. And so it’s not so straightforward to say like “Hey, NumFOCUS is out there helping to develop cures for cancer, or getting us to Mars”, or whatever the case may be, because there’s just a lot of jumps between here and there, even though that’s absolutely true.

So it’s unsexy work in some ways, right? Like, “Wow, we are helping the projects that you use and care about to figure out how to deal with this IP issue.” That’s not very exciting, but it’s important.

Right. That being said, I think you would do well to tell those stories, and maybe do case studies, even if the correlation between this particular NumFOCUS project and the end goal of getting this Rover on Mars, or whatever it happens to be, or whatever is interesting to people isn’t a dot-to-dot line… But the fact of the matter is that you are enablers; you’re enabling these people to do the work that they’re doing, and look at how important the work is that they’re doing; without these projects being supported, you’re cutting the legs off for any of these other efforts.

I think case studies, I think telling those stories, whether it’s through video, written word, or whatever medium is used, I think would do well to give the why, because at the current point you’re – and you’re in a good spot here on the Changelog, because you’re preaching to the choir; we are open source people, we understand, we donate to projects, our audience I’m sure will go to NumFOCUS.org and I hope will donate, and if you guys do, tell Gina you heard about it or found out about it on the Changelog… But to go beyond that is I think more difficult, but I think that does require some storytelling. I think that’s what you’re kind of calling for, Adam, isn’t it? Give you the why.

Yeah. Well, you mentioned yt earlier, for example. If that’s something in astrophysics, or something around oceanography, for example, what are some of the recent discoveries that have happened? Or phenomenal images of space that we’ve discovered, or how we’ve been able to go deeper into the smallest of the small, to the biggest of the big, which is what astrophysics is… You know, what are the triumphs in there that yt has played a role in. And even one step further - who are the influencers in those spaces that can then also preach the yt praises of how it’s enabled these advancements?

[01:00:16.23] Yeah, that’s great. Hopefully we’ll be able to chase down some really good examples of that. We had - I guess it was a year-and-a-half or two years ago, we found out that a lot of our tools have been used in the LIGO discovery.

Oh yeah, LIGO I believe is measuring gravitational waves.

Yeah, yeah. So we got a nice little bump from that. I’m working on a story - it’s way overdue, I’m sorry - about Anne Carpenter’s lab, which has a little tool called CellProfiler, which is used by thousands of biologists to quantitatively measure changes in cells. That work powers a lot of probably really cool stuff, like working on cures for cancer - changing the shapes of cells, as you might imagine… And her tool relies on some NumFOCUS tooling.

So some of the projects that NumFOCUS supports are deeply foundational tools; they are the base upon which other tooling is able to operate… So trying to connect those dots, and the more ideas that we can get, the better. This is another one of those challenges, it’s like, I definitely wanna go out and tell those stories, but it takes a lot of time to chase them down… So if any of your listeners are using NumFOCUS tools for something really cool and interesting that they’re doing, and they’re like “Hey, I have a great case study for you”, please get in touch, because I would love to talk to you about.

Yeah. At the same time, we’d love to highlight those things too, so as you have phenomenal case studies, or even sub-phenomenal, since I’m using too big a words for that… [laughter]

Just average ones…

Average, mediocre…

Just some okay ones… [laughter]

Right. Any case studies whatsoever…

I love that framing, “This is sub-phenomenal…” [laughs] I’m gonna start saying that during code reviews.

There you go. Any stories out there that are worth telling, I think that’s what we’d really like to shine a light on, especially as it relates to you already being on the show, and having a chance to share more of the future direction of NumFOCUS, and your achievements and progress - we’d love to be able to focus on that as well.

So if you’re listening to this and you use a NumFOCUS project, please reach out to Gina and tell her or anyone else there your story and how you’re using different projects, or the projects that they’re sponsoring, to make science possible.

Gina, we appreciate you coming on the show today and for sharing your time with us. Is there anything else on the horizon of NumFOCUS that people are less aware of, that you wanna make more aware of, as we tail off?

One of the things that I’ve been working on more or less since I joined NumFOCUS is I’m on a mission to make sure that everyone knows that the PyData event and meetup and YouTube products are NumFOCUS-organized. It’s happening less often now that people are like “Oh, what’s NumFOCUS? I love PyData”, but it still does happen.

[01:04:01.24] So now let it be known that PyData and all the wonderful, community-driven work that happens under its auspices is organized and supported by NumFOCUS. And if you go to PyData.org, you can check out our 2019 events line-up; we’ve got some confirmed dates and we’ve got some “Hold roughly this date” for this coming year, but… There should be a lot of great opportunities to submit proposals, and if you’re an experienced person or a newer person, there is a place for you in a PyData event.

We also are looking for a couple of more U.S. events this year, I believe, so if you’ve got a lead on that, or your company has space and wants to host, please reach out, because that would be great.

Let’s highlight a couple of these events then. You’ve got PyData Florence, which is in Italy… We do have a worldwide audience, so listeners in Italy, check out PyCon X; that’s May, 2019. Amsterdam, London, New Delhi… You mentioned United States, so we’ve got Los Angeles and New York City later on this year, Argentina, Berlin, and Warsaw. Quite a global presence there for PyData as well.

I think calling out the fact that when you register and pay for a PyData event to be a participant, that those funds go to support NumFOCUS.

Yeah, that’s cool.

It’s really important to mention… So if you can either help host, as Gina just mentioned, or attend, then that’s an easy way to support. Boom. It’s amazing.

Yeah. And then in exchange for attending, you can come to the NumFOCUS table and collect your stickers, which for a long time…

Now you’re talking.

…when people would say “Oh, you’re the sticker people!” and I go “Yes, let me tell you what else we do…”

Oh, the sticker people… Well, Gina, thanks again for your time today. Thank you so much for sharing your portion of the NumFOCUS story and the greater mission that you all serve. We appreciate the community that you serve, and we thank you so much for your time.

Yeah, thank you so much. It’s been great!

Changelog

Our transcripts are open source on GitHub. Improvements are welcome. 💚

0:00 / 0:00