Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Episode #26: Jim Vallandingham
140 Plays9 years ago
Very excited to welcome Jim Vallandingham to this week’s episode of the podcast. Jim is a Data Visualization and Open Web Engineer at Bocoup. He writes regularly about everything open: code, projects, collaborative tools, building great stuff. In this episode,...
The post Episode #26: Jim Vallandingham appeared first on PolicyViz.
Recommended
Transcript
Introduction to Juice Analytics and Juicebox
00:00:00
Speaker
This episode of the PolicyViz podcast is brought to you by Juice Analytics. Juice is a company behind Juicebox, a new kind of platform for presenting data. It's a platform designed to deliver easy-to-read, interactive data applications and dashboards. Juicebox turns your valuable analyses into a story for everyday decision makers. For more information on Juicebox or to schedule a demo, visit juiceanalytics.com.
Jim Vlandingham's Work Experience
00:00:34
Speaker
Welcome back to the Policy Viz Podcast. I'm your host, John Schwabisch. Very excited today to have my good friend, Jim Vlandingham, on the show all the way from the West Coast. Jim? Yes. I'm very excited to be here. Thanks so much, John. This is great. I'm glad you could get up early and come up every day.
00:00:53
Speaker
So let's dive into a few things you've been working on over the last few months. You're now at Boku? Yes. I'm an open web engineer and data visualization engineer at Boku. Boku is a consultancy group that they've been around since 2009. They do some really great work.
00:01:15
Speaker
and part of the data visualization team there that Irene Ross is leading and we are all about data analysis and visualization and interaction and we've been able to do some really cool stuff. So let's talk about some
Development of the Moibio Framework
00:01:32
Speaker
of the cool stuff.
00:01:32
Speaker
Okay. I think there's two projects I want to talk about a little bit. There's the Moibio framework, which is with Santiago Ortiz, and then there's the Voyager project. So do you want to talk about each of those a little bit and then we can dive in? That would be great. Santiago has been working on various iterations of the Moibio framework.
00:01:53
Speaker
for three or four or five years. If you've ever seen any of his work, it's almost all done out of various incantations or in varieties of this framework. But this was the time where they wanted to kind of
00:02:09
Speaker
get it prepared, get it out of his head and into a spot where other people could try it out and play with it and learn from it. So it was a really exciting experience. It was part, you know, kind of understanding the code base as it was, part cleaning up the code and organizing it and get it to a spot where other people could find stuff and then adding a lot of documentation and testing them.
00:02:34
Speaker
and support around that. So I think we did a good job with that preparation.
Features and Applications of Moibio
00:02:40
Speaker
Taking a step back, if you haven't seen the framework, it's kind of a, it's almost, it has some parallels to processing JS and also some new stuff. So it's all about data transformations, getting your data into the right type. And with that right type, then you can start
00:03:02
Speaker
So it has kind of a canvas-based drawing library built on top of the data types that you can create. Does it allow you to manipulate the data within the framework itself?
00:03:13
Speaker
Yeah, so typically, if you start out with an array, you can turn that into a list or a table. And they kind of work a little bit like stuff that you might find in R or Python's pandas package, wherein you operate on the table kind of in a vectorized fashion or the list more than once you have the array in that setup.
00:03:41
Speaker
What's the value or what does this framework give people that JavaScript or processing or some of the other tools don't? Yeah. Well, it's written in JavaScript. So it's not, I don't think it's a competitor. It's more of like a system that has worked well for Santiago and so could work well for other people. Gotcha.
00:04:05
Speaker
It does do a lot of stuff around speed, so making sure that the rendering to Canvas is really fast. It also has a very interesting model around interaction. So if you've ever worked with Canvas stuff, interaction, knowing what you're clicking on and where the mouse is can be a problem. The tooling around the mobile framework allows you to
00:04:31
Speaker
get a return value from like when you draw a circle then you can check to see if this the mouse is over that circle at that at that point in time and then deal with or you know change change the visualization based on that that knowledge right so it's all the kind of baked into the
00:04:49
Speaker
the drawing primitives themselves as opposed to a separate system. I see. So do you view it as a system that anyone can sort of jump into and use or is it sort of more for this more sophisticated user or the user who's trying
Community Building and Contributions to Moibio
00:05:05
Speaker
to build sort of more customized things the way Santiago
00:05:08
Speaker
Yeah, I think it's for, I don't think it's necessarily going to appeal to everybody. I think this is the foundational groundwork for some much bigger plans that they have, which he alluded to at the OpenVizConf, the Lycan project, which is using the mobile framework to create
00:05:28
Speaker
a completely visual non-programming kind of environment where you connect your data and you connect your visualization kind of by a drag and drop method, which is really cool and exciting, but this is kind of the underlying foundation that helps power some of that capability.
00:05:45
Speaker
But it's now available separate from the Lycan project. The Lycan project is not quite out yet as a public entity. But with this, you can start trying out some of the interactions, some of the conversions that Mobile Framework offers.
00:06:05
Speaker
It's on GitHub. When Boku or when you put things on GitHub, this type of project, for example, on GitHub, are you trying to work with people to further develop? Or is it just like, here it is, go off on your own? Or are you really trying to interact with people and develop it further? That was a big part of this engagement, was how to start building that community around this.
00:06:30
Speaker
It's a surprisingly tough problem to tackle. And so it took a lot of work around getting the infrastructure there, getting the support there, crossing the T's and such about who was going to be supporting this after Boku left it officially. But yeah, we're still there. We're watching the
00:06:54
Speaker
the issues in the uh... that kind of contributions from uh... externally i think it's going to get a little attraction thought is a lot about uh... getting documentation for the examples of you know we're
00:07:06
Speaker
certainly inspired by the success of D3 in terms of how the paradigm of a large body of examples can work wonders for. So we're nowhere near in terms of the number of examples or our demos, but we have kind of the start of that as a showing of what's possible with the framework. And we have a cool demo reel, which everybody likes.
00:07:31
Speaker
Yeah, that was good music. I enjoyed it. I was down in my head. That's GarageBand for you. There it is. You can make whatever you need.
The Role of Git and GitHub in Collaboration
00:07:40
Speaker
So I want to talk about Voyager, but let me sidetrack for a second and talk about Git and talk about GitHub, because you have a really nice set of posts on how to use Git, I think. So what was the inspiration for writing that set of posts? Well, this was in a lot of my
00:08:00
Speaker
work, I've been promoting the use of collaborative tools like Git and GitHub to various degrees of success. And so we saw another opportunity to provide kind of an example workflow that could be used for small groups that are perhaps familiar with the basics of Git or have dabbled in it a little bit.
00:08:29
Speaker
As an individual but need to now figure out how to work together without going crazy And so this yeah, this series of posts kind of is derived from some of the work that we did with Santiago and just kind of Codifying it into a bunch of blog posts that that go over
00:08:53
Speaker
Just anyway you know that's the problem with some of these technologies is there's a thousand ways to do something you don't know if you're doing it the right way and there is no right way but there are suggestions or processes that you can use that that hopefully make it so that people actually stick with it you know right.
00:09:11
Speaker
where the benefits outweigh the complexities involved in incorporating something as sophisticated as Git into your day-to-day workflow. Because yeah, it's not necessarily straightforward. It's not easy when you're first getting started with Git or any collaborative tool and figuring out how to use it
00:09:36
Speaker
with the least amount of friction without having too much magic so that when something more problems arise cuz they will yep they should be old know what's going on enough that you can figure it out yeah there's a lot of lingo that i see a lot people sort of have this was like this beer and to serve understand the lingo the other vocabulary is one yeah one that we wanted to start breaking down yeah and if you know it's it's a worker you know it's i think we do okay job like uh...
00:10:03
Speaker
feature branches, the term gets thrown around a lot. If you're coming in this without knowing what that is, then you're going to get stopped there. The start of this is attempting to demystify some of that and bring it down to
00:10:21
Speaker
What are we actually talking about? It's just a branch and going over what branches are in Git and how they're useful and why they're useful. Now you've worked in a variety of different sectors really. You've worked in like the healthcare sort of sector, you've worked in the corporate sector, now sort of the open, this open data sector with POCU. So when it comes to these collaboration tools, what are the differences you've seen from people in those different sectors?
00:10:51
Speaker
Well, you know, I think it's I don't know if it's a sector based so much as just an individual So, you know, I think kind of a common theme and a lot of the the environments that I've worked in is that there isn't a strict hierarchical authority coming down from above at least in terms of Workflows and and processes that get used from the technical side. So Most places aren't saying you have to use get
00:11:21
Speaker
and you had to use it this way and XYZ. I think those environments are definitely there, and maybe that's the more common, but in my experiences, it's been more around. We have a group of individuals with various backgrounds and experiences, and they need to learn to work together. We need to learn to work together, and so I've had
00:11:44
Speaker
people who don't want to use Git at all, they don't see the benefit and enjoy the version one, version two, version three copy and paste system. And so trying to typically unsuccessfully motivate them to switch around. And then we got other people that have experienced other version control systems, for example, SVM and want to use
00:12:13
Speaker
don't see the benefit of moving from their tools or want to use the new tools like the old tools and that kind of stuff. Which is all valid things and that's the give and take of these kinds of tools.
00:12:31
Speaker
They are useful, but also complicated and can add additional friction to your processes if you don't know what's going on. Right. And there's an upfront cost that people may not want to learn. Learning stuff. Learning is hard. Okay, so let's talk about some tools.
Exploring Data with the Voyager Project
00:12:51
Speaker
Let's talk about Voyager, another project you guys did with Jeff Hare at Washington.
00:12:58
Speaker
and ham and and um... and and you know to be fair they've they did not almost all the work you know we're not we're not trying to swoop in and say hey this is our project right there this is a great existing academic uh... work that was done uh... in hairs lab uh... by some very talented developers and and visuals like visualization experts in their own field right uh... we were fortunate to be able to work with them with uh... the help from a night uh... uh... uh... prototyping uh... grant right
00:13:28
Speaker
And we kind of came in with the idea that we knew there was something very interesting about this tool, but we wanted to improve it in ways so that new users or people not from the academic world would be more familiar, more likely to try it out. And also that if we could improve the speed and the size of the data sets that it could handle, then that would be kind of a win-win.
00:13:56
Speaker
And so Voyager comes at the idea that it combines a set of best practices for visualizing data with a recommendation system, which is kind of a very interesting concept, where you throw in your data set and it starts generating
00:14:15
Speaker
recommended visualizations with the idea that you're trying to see as much of the data at one time as possible. So you don't get stuck in a local minima of exploring, but you rather allow it to show you kind of the breadth and the width of the data
00:14:36
Speaker
but still clueing it into onto what you think is important. So you can click on a particular column in your data set and it'll start regenerating more graphs that include that column plus other columns and transformations of those other columns.
00:14:54
Speaker
So the demo, when you start it up, I'm sure you'll add the link. I sure will. Yeah. When you start it up, you get kind of a classic dataset around cars and the number of cylinders, their names, where they come from, the year that they're made. And you can start exploring it right then. So if you were to click on number of cylinders, you get
00:15:20
Speaker
variations of the graphs that include cylinders, but also like acceleration or the average acceleration, horsepower, and so on and so forth. So you can start to see correlations. You can start to see outliers and stuff like that. So it's really cool. You get a sweeping view of your data set organized and kind of motivated through this recommendation system.
00:15:50
Speaker
So I'm always curious about these recommendation systems because there is somewhat of a best practice idea and date of it, although there's always a pushback against pie charts are good, pie charts are bad, that sort of thing. So do you have a thoughts on when you build these recommendation tools about
00:16:11
Speaker
What should be a part of of that library and what shouldn't be a part of you just say look are the job here is just to let you Explore we're gonna get right whole library for good or for worse. Well, I think I think in this in this case I didn't I think the domain is constrained enough that it makes sense where they're they're not talking about at this point in the in the in the process of dealing with data the time that you would go to Voyager or
00:16:37
Speaker
You're not looking to create an explanatory graph. You're on the pure explorer side. And so with that constraint and with the constraint of keeping everything kind of consistent so that you can use very common charts to explore a wide range and diversity of data types and data transformations.
00:17:00
Speaker
I think it makes sense to have this small set of graphics that you can continue to work with. I know when I do data analysis, typically I'm still in R, and I'm typically using histograms and box plots and bar charts as the primary tools, and scatter plots obviously, of exploration.
00:17:27
Speaker
And that's because those tools work so well for everything. Now, when it comes to time to generate explanations or presentations around this data, then that's when you would start deviating from the best practices and going with what makes the most sense to tell the story that you want to the audience that you're interested in.
00:17:54
Speaker
But yeah, I think automation of this kind of stuff is valuable and can only exist with some best practices in place and adhered to, I guess. Yeah. Very cool.
Insights from the OpenVizConf
00:18:08
Speaker
Before we close up, I want to talk about OpenVizConf.
00:18:11
Speaker
which is coming up in April. Your number four quickly has become sort of one of the important Dataviz conferences out there, right? I hope so. Yeah, it's supported by Boku and Irene and Lynn Cherney are both co-chairs of it and it's been really exciting.
00:18:30
Speaker
I've had the privilege to be able to speak at two of them, the first one and the third one. And yeah, it's a really exciting time. They have the technical sides, they have the design side, they have the academic side. We're excited about the speaker lists, the proposals that we've just got in. I get to be a committee member now. I haven't been to co-chair yet.
00:18:59
Speaker
No, I don't plan. One has to have goals, Jim. I can dream. So it's been really exciting to see it from this side of the spectrum, too. And it's going to be at the Boston Aquarium. So if you go to the website, openvizconf.com, you get to see jellyfish and a bunch of aquatic stuff.
00:19:25
Speaker
Yeah, it's going to be great. That's a pretty incredible place to have it. Do you know offhand how many applications or how many submissions came in this year? I think it was well over 100. It's interesting the amount and the spectrum of which the talks were covering. So it's going to be tough to whittle it down. But I think that's just a win for conference goers now.
00:19:53
Speaker
I think we're going to have a really good lineup. We're looking at the visualization side and also a bit of the data side and a bit of the machine learning or how those things are combined in the world of visualization.
00:20:10
Speaker
It's pretty exciting stuff. That's great. That's great. Well, I'm looking forward to it. I'll be out there. Get your ticket. I got my ticket. I'm all set to go. John's going to be there, everybody. So you should be there, too. Yeah, I should be there, too. And I'll put links up to OpenViz and all the other projects that we talked about today. Awesome. Jim, thanks so much for coming on the show. It's been great. All right. Thank you very much, John.
00:20:30
Speaker
And thanks to all for listening. If you have any comments or suggestions on the show, please let me know in the comments on the site or on Twitter. And please rate the show on iTunes or your favorite podcast provider. So until next time, thanks for listening. This has been the policy of his podcast.
00:20:56
Speaker
Again, thanks to my sponsor, Juice Analytics. For 10 years, Juice has been helping clients like Aetna, the Virginia Chamber of Commerce, Notre Dame University, and US News and World Report create beautiful, easy to understand visualizations. Be sure to learn more about Juicebox, a new kind of platform for presenting data at juiceanalytics.com. And be sure to check out their book, Data Fluency, now found on Amazon.