Tonya Filz | The resilient R champion | RStudio (2019)
Merriam-Webster defines resilience as the ability to recover from or adjust easily to misfortune or change. As a Customer Success Representative who works alongside data scientists using RStudio’s toolchain, I’ve had a front row seat to the challenges faced by data scientists as they aim to promote the use of RStudio’s toolchain in their organization. This talk will focus on effective strategies that have been used to overcome some of the most difficult organizational barriers that are faced by data scientists using R. Specific topics will include funding barriers, IT support, server space, the “open source mentality”, and political pressures within organizations. About the Author Tonya Filz Helping data scientists and data analysts succeed by leveraging RStudio to clean, visualize, analyze, and communicate conclusions to key business leaders
image: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
So my name is Tanya. I've been with RStudio for about two years now. So I work in customer success, so I work with several of our enterprise customers. My background is as a therapist, so that was my first career. And so I started to see these principles of resiliency in some of my strongest R champions.
So I wanted to take some time to talk through what an R champion is, how you can improve your R championing. Is that a word? I think it is. Yeah. R championing and some of the barriers and some of the ways that I've seen folks be successful in overcoming those barriers.
Who is an R champion?
So who is an R champion? Well, probably a lot of you here are R champions. But these are some ideas that I came up with as I was thinking about what an R champion is. So an R champion is someone who uses R to accomplish the core goals and objectives of their job, promotes the use of R within their group and broader ecosystem that they live in. They contribute to the R community through RStudio community and Stack Overflow and other mediums, advocates for and takes ownership of R tooling in their environment.
So one of the things that I'll talk a little bit more about later on is that sometimes what I see is that R is installed and left to IT to manage. And the teams that I see be the most successful are when a data scientist is really driving their R tool chain forward.
And then an R champion is also someone who stays up to date on all things R and freely disseminates this information. So within their group at work, but also through other mediums. So through community, through R meetups and other channels.
Simple ideas to improve your R champion cred
So I wanted to add in this slide for simple ideas to improve your R champion cred. So one easy one is to organize a lunch and learn for just R analysts or maybe analysts interested in R in your group. And then once you get this off the ground with your friends, invite more friends from parallel groups. So how do you get a hold of them? Start an internal Slack channel. So for folks in your broader organization that are using R.
So I saw this be really successful in one of my accounts where there were several different branches that were all technically different orgs but under one broader org. And they started a Slack channel for all the different branches even though they were technically different companies under the same parent company. And it was really successful because when someone ran into a problem with their R code, they could rely on their broader colleagues in their bigger ecosystem for help.
Be intentional about connecting with other R users at R meetups and R conferences and talk about the barriers that you're running into. Because chances are you're not alone. When you run into a barrier or someone says no to something that you want to do, chances are it can feel very isolated, but chances are you're not alone. Someone else has run into that. So talk about, you know, the 50,000-foot view of how things are going in your organizations with other R users.
Barriers R champions face
So what are some of the barriers? Now that I got you all stoked about being an R champion, what are some of the barriers that you might run into? So these are definitely not all of the barriers, but this is a short talk, so I can't go through all of them. So some of the ones that I see are the open source mentality. So that was kind of twofold. Why should we trust open source tools and why should we pay for open source tools? Funding issues, there's not enough money, IT support, and then, of course, internal political pressure, everyone's favorite.
So why trust open source R? So one of the things that I've learned a lot about in my role is the different mentality. So the old school thought before open source became utilized in enterprise settings was that you buy your software from a big company that wrote the software, and then you have to pay for support to support that software. And I don't know if you've ever bought a software from a big company and then had it not function as you anticipated. And it's hard to really validate that because it's closed source, right? You cannot see the source code. You can't validate for yourself that the software will perform as anticipated.
So active software utilization is more the open source mentality. The code is all there, and you have to or you can go in, look at the software, look at the code, and make sure it will do what you need it to do. And I will say it's a little bit higher investment with the active software utilization. So that's sometimes where you get some resistance is the fact that you're taking ownership of it and you're validating it yourself.
So the nice thing about that, though, is open source technology, it's really hard to hide faulty software. There's kind of nowhere to hide. Your source code is out there. If it doesn't work, you're going to find out really quickly and the community is not going to be happy. So that's another big benefit is that you can validate your own software. It is a higher investment, though.
So the nice thing about that, though, is open source technology, it's really hard to hide faulty software. There's kind of nowhere to hide. Your source code is out there. If it doesn't work, you're going to find out really quickly and the community is not going to be happy.
So there's a strong community for our developers and our platform administrators. So we're doing more and more on RStudio community to support the IT folks who are supporting the platform for analysts. And last but not least, sometimes you can run into issues with your legal team or your info security team. And there is commercially licensed R and tools for R available, and sometimes that can help assuage some of the legal concerns when you're switching between or when you're moving more towards an open source software.
Why pay for professional tooling?
So okay, now that you've got the open source software, why should we pay for tooling around that open source software? There's a few different reasons. So performance, so I don't know if you've had any experience with the open source tools versus the professional tools, but typically you have a little bit better performance definitely with the server products that we have. You have improved performance. It can be deployed in a more secure fashion. You can have things like authentication, kind of locking it down, and that helps InfoSec feel more confident in the tools. And then you can have increased collaboration with your colleagues.
And then with Connect and other tools, you can do some automations, you can automate out some reporting, you can also have a commercial license, which we kind of talked about before, but you can have some legal coverage, which makes legal happier. You also get maintenance and support, and someone like me and support and solutions engineering to give you ideas about how you might want to set up the software and give you suggestions. So that can be really valuable as well.
And then there was this idea that Tarif talked about in his keynote yesterday of technology philanthropy. So by investing as an enterprise in enterprise toolchain for R, we're also helping the broader R community and the RStudio community as well.
Overcoming funding barriers
So this is another one that I hear a lot, funding. There's not enough room in the budget. How many people have run into funding barriers? Yeah. It's really common. This is probably the top one that I hear.
So my best advice for overcoming the funding issues is just to use the open source toolchain. So really deep dive on creating APIs, creating really valuable shiny apps. Show them what's possible. And then be able to communicate what else is possible with the toolchain. So I like folks to start with the open source toolchain and create a demand for a parameterized R markdown. So really deep dive into the open source toolchain. And a lot is actually possible with the open source tools.
So collaborate with parallel teams through your R community. So once you've built up an R community, collaborate with those teams. I've seen several teams be successful in collaborating and resource sharing. So two different teams and different organizations using one software. That's definitely possible. Or if there is a team in your organization that's been successful with funding, they may know how they may have ideas for you as to how they got that funding made available to them.
And then I kind of covered this in my first point. Clearly communicate the possibilities with open source and the possibilities with professional software. So once you have your proof of concept, there are evaluations of all of our professional products that can allow you the opportunity to demonstrate what else is possible. And once you create that appetite, I've seen funds become available.
Ask good questions. So I think sometimes when we're told no, the automatic response is, oh. And it can be hard to kind of pull yourself together and approach that with curiosity. Oh, okay. Is there any funding changes in the future that you can see coming down the road? What about our tool chain for next year? Is there going to be any additional funding? So sometimes just getting over the, oh, and going to approaching it with curiosity can be successful in kind of problem solving.
And the way I like to think of it is help your manager or whoever told you no solve your problem. So you want to get them bought in. Ask a lot of how, what, open-ended questions in a very curious manner because you're really trying to do them a service at this point when you're asking for tooling. No one just asks for tools just to have a bigger toolbox.
Navigating IT support
The other thing that I hear occasionally is from IT, we can't set up something that is free. There needs to be a commitment from the business. Again, asking good questions becomes really, really important here. So what commitment is necessary to use open source tools? What does commitment look like? Can you give me an idea of what that looks like? Is there a separate team that manages open source software?
So I've had several of my larger accounts that I will work with. There's the IT team that they always use, and that team says no, and then they don't really know where else to go. But sometimes within your organization, there's a research group that's kind of like a mad science lab where they're deploying different software, and sometimes those mad scientist groups are way more willing to take a chance on open source.
And another thing to think about is, you know, they're saying no to R, but I can almost promise you there's Python somewhere. What team is managing Python? Like is there a team for managing open source software? So sometimes it requires a little out of the box thinking to find the right team.
And asking about the specific policies, because sometimes they'll say, no, that's not in line with our policy. Well, okay. So sometimes just, you know, going back and asking a good what question can be really helpful. There's also a few different options from a technical perspective, so we do have pre-configured virtual machines on all of the major cloud platforms, so AWS, Google Cloud, I think Azure as of recently.
And then, yeah, find those open source policy documents. And then the last thing is explore external partners. So if you're really kind of spinning your wheels, you can't find anyone in your IT group or your mad scientist IT group that wants to set this up, then we can help you find a partner that can come in and set up the server, and sometimes I have several folks that will need to do that just to get anything done.
Dealing with political pressure for other tools
So political pressure for other tools and languages. So not every tool is right for every problem. Don't argue for one tool. Instead argue for more tools. Again, I like this idea of curiosity, and I've seen it be really successful.
So asking about the features that they saw in the product, like what is their tangible takeaway that they're interested in? Are they interested in a dashboard? Are they interested in a report? What is the takeaway? And then in the back of your head, be thinking about could I show them something similar with the open source R? So just asking those good questions so that you understand what their needs are, and then you have your R skills that you can think about how to create that with the tool chain that you prefer.
So how does this align with our goals and needs? So asking about how the software, what direction are we going as a team? This can be really good for helping you navigate through your career. What other tools are available for me? So if I run into problems, what are my other options? So again, you're kind of recruiting them to help them solve your problem.
And so just through curiosity and letting them, telling them kind of what's possible. And then more tools is not necessarily bad. You can definitely have too many tools, but a toolbox with just a bunch of hammers is not that helpful. In this photo, it would be helpful to have a screwdriver.
Last but not least, you can do it. I've seen a lot of people run into some really stubborn barriers, and I've seen these concepts work in action. So you can do it. You can advocate for your tool chain, and you can be successful using the tools that you prefer to use.
Q&A
Yeah, definitely. So I think RStudio Connect is a great way because of the emailing feature. So every executive that's consuming your reports is attached to their inbox. And so if you can show them the automated reporting where they can get their daily reports sent to them at the same time every day, and that can include metadata in the subject title, that's a golden way to get buy-in in RStudio Connect. You can even do some really nice HTML email formatting so that it looks really great in their inbox. And I think if more executives were able to see that, they would get more excited about the RStudio professional tool chain, specifically Connect.
Do you have any experience working with public sector agencies where, I mean, I'm from Washington State in the Department of Health, and Microsoft is kind of a religion there. And so we have contracts with all these things, and trying to slip this stuff in is more than challenging.
Yeah, definitely. So we have several folks. So actually, the team that I worked with that had this giant Slack channel, they were actually in the public sector. So the other thing from a legal contracts perspective, you probably know purchasing anything with the government, I mean, you need like so much, there's so much red tape. But we're doing more in terms of contracts, partnering with folks that actually are on the GSA schedule and other things that public sector and government organizations really need. It is possible. There are more barriers and it generally takes longer, but we're doing more and more to make it easier to purchase from us.
