EP10 – The Tradeoffs of Deploying Open-Source Vs Commercial Geospatial Technology

In this episode of On Point with Korem, I had the opportunity to speak with Andy Dearing, the former CEO of Boundless Geo and now General Partner and Geospatial Project Lead at GeoFutures STL, an organization leading the economic development of geospatial technology in the St. Louis area. We spoke about a breadth of topics including the changing nature of deploying both open source as well as commercial geospatial solutions and the tradeoffs of both and how this has impacted the ability to find qualified technologists that are skilled in both programming and location analytics.


Joe Francica : In this episode of On Point with Korem I had the opportunity to speak with Andy Dearing the former CEO of Boundless Geo and now a general partner and geospatial project lead at Geo Futures Stl. An organization leading the economic development of geospatial technology in the St. Louis, Missouri area. We spoke about a breadth of issues including the changing nature of deploying both open source and commercial geospatial solutions and the trade-offs of both and how this has impacted the ability to find qualified technologists that are skilled in both programming and location analytics. Stay tuned for On Point.


Joe Francica: Andy, thanks again for agreeing to do the podcast. Really appreciate it, appreciate your time. Let’s go back a few years to your time at Boundless and look at the state of open source geospatial technology and it doesn’t seem to be like an either or case where somebody’s either going fully open source or fully commercial. What’s your take on kind of the mixed geospatial environment that we’re seeing being deployed these days?

Andy Dearing: That’s a great question. What we’re seeing is exactly what you’re saying, most customers have some sort of a mixed environment. A lot of organizations you still see that were traditionally using commercial software you know and non-open source and if they need to have a specific solution deployed, but a lot of those customers especially ones that are dealing with heavier data sources or maybe a veracity of the data coming in is pretty…and they need to have different technologies to be able to support that. You’re seeing that blend and I think with technology moving as fast as it is with regards to data storage, cloud versus on premises, they just need something that works and so finding that right mix of technologies that might scale with what they need but also have that solution piece at the end, that’s where it really lends itself to that mixed environment of saying “Hey, you know what I might want to use open source on the back end to manage the data or to scale and bring that there, but my customers might want something on the front end that integrates in a certain way” and so I’ve definitely seen that shift and it’s been in parallel with the amount of data that’s being produced and consumed from the back end and what the customers are really looking to see on the front end.

JF: Does that kind of present a push-pull situation where there’s needed expertise in one area and maybe not in some areas. Like either, you’ve got expertise in open source, or you’ve got expertise in big data and is that putting a strain on the workforce of geospatial today?

AD:  Well the workforce as you’re aware Joe is completely changed in what we would consider geospatial. The workforce is not…we started seeing a change where you would have the IT department and then you’d have the GIS department and they would have to kind of work together and then as things morphed along you started seeing a geospatial developer for example, right. Where they could actually program and make some maps or do some data integration work but they also knew GIS concepts. Nowadays, you have technology companies and groups and organizations that have just a bunch of developers or data scientists, that may not know anything about geospatial at all and then they’re expected to but they’re using it as a data source to solve some problem or to produce some product, and what they’re used to or what they’re trained with is it might be just IT tools or it might be they’re using R for example, and using some programming around that they’re not using you know some tool boxes that are already out of the box, they’re programming in a certain language but want to bring that data in. So I think the complexion of the workforce changing into it being more, I would say tech workforce, and how they’re dealing with these data streams, is kind of forcing that and they’re expecting something different when they land into the job or they’re expecting to be working with, maybe something like an open source tools that’s how they work that’s how they do their work, and then they kind of learn or see what tools are out there that might make their job easier. If there’s a commercial tool that makes their job easier, that doesn’t break what they’ve done, then they’ll use it. But I think it’s just that the experience coming in and the complexion changes of the workforce are kind of the modern geospatial workforce that we have today.

JF: Yeah, and I had a friend who I worked with many years ago, who was kind of a just a geospatial engineer. He was kind of a demo jock. But, he went to work at the city of Santa Clara as the GIS manager, well he’s no longer the GIS manager, he’s the IT manager and he’s looking not just for GIS people, he’s looking for data scientists. And that’s a huge change that I see, I guess you see the same thing?

AD: It was interesting as we’ve been doing some reports of analysis here locally on the workforce number of jobs, we were starting to look at some job openings and so you’d see some of these job openings changing at the big government agency here in St. Louis and that they started having data science jobs or data analyst jobs. You’re like, well that’s different right, it’s not the GIS analyst one or two that they typically used to have but then all of a sudden you started looking at health institutions and you started looking at the requirements of those positions being a data scientist or data analyst and one-to-one the skill sets of the requirements skills that are needed to do that job are almost in parallel ,they’re almost like IT jobs, and then when you get to visualization tools, we actually solve the health systems, they’re like “oh and experience with tools like ESRI and Tableau” and you’re like wait! Like we would never have seen that this is a geospatial job but by the way this data scientist has to have that tool in their toolbox to produce some maps that might have some of this information on it. So it’s a different shift that I think we’re all kind of seeing there. I think in the next five to ten years they’re going to be tech jobs. That’s really what the future geospatial workforce is going to be.

JF: You know I had the opportunity several years ago to work with the department of labor on their geospatial competency model and, of course, everything was very geo-oriented and now I wonder just from your comments whether we’ve got to change that competency model to include more programming, more you know experience with R, or java, or Python?

AD: Yeah, because well you used to have the DBA back in the day… you’d have the IT person that would hook up the wires and make sure the servers worked, you’d have the DBA that would take the data and put it together, then you would have an analyst that would go and analyze the data and put it together, do the geospatial stuff, and then you’d have a developer and now you’re expected to have all of those into your back pocket in some way shape or form.

JF: Yeah, so that brings up the another interesting situation where we’ve got new suppliers of technology out there. You see Google with Google Big Query, and you see Snowflake and Data Bricks, they’re all now, I don’t know forced into supporting geospatial data and technology, I mean that’s a huge change, that’s a big step into our world, right!

AD: It is! It’s good to see that. We have to get them to realize sometimes that “wait, location data is actually important. You are kind of a location geospatial company that you don’t even realize it, right! But, also getting them to realize that…we always hear the adage and I always kind of go back and forth that geospatial is not “geospecial” but in some aspects, there’s some aspects of the science that we come from, that you need to have an appreciation or understand some of those models or understand the implications of it, because if you look at data just as data or you look at information as information and don’t understand the science behind what we do, then you end up having false information or wrong information that gets put together or you’re doing something like “why are you trying to do it that way, there’s tools out there that can help support you do what you’re trying to do” and so that’s where I think this mesh of… and I bring it to like the satellite companies or other organizations that are producing large volumes of information, the IT paradigms have to change, and so these you know the organizations or technologies that you’re talking about is mainly about either storing or managing or serving large streams and volumes of data, that’s not going to stop anytime soon. And then you have people that are analyzing that information anymore you’re building algorithms to extract the good tidbits out of that, right. And, so now that locational or coordinates or those sorts of things are sitting in this data and then you’re wanting to actually maybe put it onto a map or do that analysis to do that, then that’s where they’re like, “oh wait, I don’t know how to deal with coordinates,” or “wait, what’s a projection model and why do I need to figure that piece out”? And you’re like, “well, based on the end result, we need to have you store it this way or serve it out this way or analyze it in this way. So, it’s good to see that problem finally coming together because I think that’s going to light up more geospatial opportunities that are out there, that have never been exposed before. Hopefully.

JF: Well, so let’s take a slightly different approach then. Is it necessary to teach somebody how to think spatially? So, it’s one thing to process the data, it’s quite another to teach somebody how to construct a spatial query in such a way that they’re deriving meaningful information from that data and I wonder if that’s a special skill? I used to work for a company that their tagline was “Thinking Spatially.” And I don’t know whether that’s a specialized skill or whether it’s just something we do maybe more naturally today.

AD: Well and I think hopefully we don’t have to do those complex things like we used to have to do that, hey they’re already built into certain things like data transformations or other things that are out of the box today that used to be very complex in the past, to where you did need a GIS certificate to understand what was going on. Maybe that becomes less intense but I think to your point of understanding the implications of why and having an appreciation for what that means and as you’re producing your end result work. Maybe that’s where it is and that you understand or you’re maybe you’re thinking having a spatial thought around that or understanding the importance of what that is that goes into the end product that you’re developing. But we would have a lot of developers or we would meet a lot of customers at Boundless, even before when I was at ESRI, where they had a lot of high-end developers that were just jamming some code away and they would always come to us because they were having a problem, right. “I can’t get the data to do this, whatever” and you’re like “well, did you think about this” and it was a spatial function or something like that and they’re like “no, I didn’t know that” or “oh, I didn’t realize there’s a toolbox out there to do that, that’s really complex.” We’re like “well no, that’s just table stakes in the business that we’re in.” We know this and so it’s whatever that part is bridging that gap with them, it doesn’t have to be super complex you don’t need to have the GIS master’s degree to do that but at least understanding what that is and kind of the data that you’re working with and just the nuances of what’s going on there.

JF: So, it kind of in this crossover between someone who may be coming at it from just a programmer point of view or coming at her from a GIS programmer, a GIS specialist, if I ask a typical spatial query like; “find me all the class, an office space between 4th and main in Peoria, Illinois between 2 p.m. and 4 p.m.,” do most people know how to construct an SQL statement like that without understanding the benefits of spatial information?

AD: That’s a fair point, that’s a fair point! Probably not, they’re probably like, “Man, let me google that, let’s see what’s out there,” right and then go back to the open source piece. I think the cool part is, there is actually a lot of information out there on how to construct that on your own. Are they going to make mistakes with it? Yes! But, I think to your point, no! They’re probably going to know how to do that and unless it’s already built into whatever they’re using they’re going to have to create that. They’re going to be finding a lot of articles around there. We use it all the time just like, “hey, I found this piece of open source code that does that work”? Or like, “it does but do you know what it actually is doing? Nope, don’t know.” It like okay let’s take a couple steps back.

JF: Yeah, it’s an interesting problem because it’s one of those things where maybe if you’re a GIS person you learn this but if I’m a regular IT person, I may not have thought of it in quite those terms. So, I still think there’s a bit of an education process that maybe we need to offer to some developers but it’s an interesting challenge.

AD: Yeah and it’s that meeting in the middle because I think us geospatial people we can also be pretty suddenly saying; “well, that’s really, you know that you need to have a degree on this. You need to have an appreciation for this. You need to be a geospatial person. And I think some of that we have to let go of some of that saying; hey, there are capabilities that are out there that can do this or normalize this for the common developer or user. But likewise, I think the developers also need to come back in, or you know traditional computer science or data scientists, and come back and say; “hey, this stuff is a little bit different and I need to learn more about that.” So, I think there’s going to be a meeting in the middle between the hardcore traditional GIS people and the tech community and hopefully that’s where the next levels of innovation are going to occur.

JF:  Yeah, so let me take you to another part of your expertise, which is in on the investment side. And what I saw recently you probably saw too, Planets going public via SPAC. Interesting development. I think they finally realize they’re in the data business and not the satellite business. So, I’ve just been amazed at the amount of money pouring into geospatial, maybe more so than the last 10 years. What are you seeing? Is that new development or what’s going on?

AD: You know, you’re spot on Joe. The space is like the new frontier again, right. And, I think some of that is, I mean, we’ve been doing that for a while. You’ve seen organizations like Max R, Consolidated Grow and you’ve seen…it’s been happening over years, but I think people, to your point, are understanding it’s not…it’s data and the value of this digital commodity that’s out there and how much of it can you get, is a big interest. And they happen to be a geospatial company, right…they’re taking satellite imagery but they’re applying it to a bunch of different things. Spire is another great example and you’re going to see more and more of these companies that are sensor companies that are out there. I mean, there’s more seed investment going into these companies now more than ever. We just had a company here in the Midwest that won the Arch Grants and some early C-stage capital, and they are a balloon company, and they do persistent kind of monitoring or image capturing over certain areas and they operate in a different sphere than the satellites do. But, it’s just, it’s interesting, right! As you’re starting to see companies and people investing in that because they realize the value of the information product, and somebody is going to see value in that and there’s probably a product or data behind that can be produced. So, I think it’s interesting, though, is from the software product standpoint, in those types of companies and those types of start-ups, they’re not…you have Map Box for example, but Map Box is also a data company to many aspects, right. They’re dealing with massive amounts of data. That’s where their value is. And, they have platforms for visualization, everything else, but I think it’s going to be more interesting is, what does the software business look like and is it going to follow the same trajectory of investment or is that more just a tech play versus a data play. So, space is definitely hot right now and I can foresee it continues to be hot with investment and opportunity, and people are bolting on different new sensors every day and I think it’s not as complex or hard as it used to be to get your satellite into orbit, right, and actually get it up there. There’s basically the Uber systems of getting those things up there. So, I think it’s right for innovation. My big question is; will people be able to get the value out of that data downstream, right? It’s great to capture all this data and it’ll be valuable but are people really going to get the value? Are they going to make better decisions? We’re going to understand the planet better? Are we going to make more informed new innovative products downstream with that information? Time will tell on that. Hopefully, we see that happen.

JF: Well, I think that goes to the heart of the issue, which is, I’ve often thought that there’s way too many pixels out there, right, and not enough information. So, are these companies; a Max or a Planet, going to understand that there has to be derivative data, derivative attribution, from this in order to make it useful and then monetize that data to create products that are capable of being consumed in whatever it…BI’s tools, GIS tools, whatever and I don’t know, I don’t see that coming immediately. I see it down the road and I don’t know what that’s going to take. Whether it’s a better realization that there has to be some financial benefit and I don’t know if I see that yet.

AD: Well, I always like listening or reading to Joe Morrison’s articles that he puts out there on his blog but that that was kind of one of his more recent ones where he said, “you know, there’s obviously value and people are creating value on the data side of things, and there’s value on the solutions, right. Developing an algorithm to sift through that data is not enough, right, you’ve got to get to the real value thing. How is it going to help me identify something new? Get something, somewhere, faster? Or, make something more profitable? Or whatever those business results or end results that need to be on the downstream end of things. Because it’s not good enough and I think even that will come back to the data providers, it’s not down the road just dumping a bunch of data on the loading dock every single day and just leaving and saying, “hey, good luck.” Customers are gonna want something more. They’re gonna say, “man, I can’t deal with this amount of data. I can’t deal with this information. Even though I have all these algorithms running, all I want is the answer to the question I’m asking out the data. I don’t want all those pixels, right. So, I think, unless you’re a big government agency where you might have multiple uses for those pixels or you want to see a large change over time or certain things like that where you can see value in the pixels and that kind of velocity of them, I don’t know, I think that’s what’s to be reckoned with. Beyond just supporting natural disasters or having assets that are there at the time, you need them. What about all that other data? What does that really mean? And, are you able to really give value to your customers? Is that really what they want?

JF:  Yeah, I think you again you hit the nail on the head there with, “people want answers, they don’t want data.” And what I wonder is; are we going to have to put more emphasis on machine learning tools to do that for us? And, is that a good solution, right? I mean, is that the answer or I mean, if you become too reliable in machine learning then you also miss something as well in the analytics.

AD: Well, going back to the workforce question; do those people that are writing those algorithms actually know what they’re looking for. Do they know that, “oh wait, if you hit this different color band and you interrogate it the certain way, that this is the result you can have”? And that’s actually where spatial really comes right back into the fold from a remote sensing perspective. Where having those remote sensing skills and understanding of what that data actually means. You need to start arming those analysts with that because otherwise throwing data analysts or data scientists at these things and developing algorithms, they’re still not producing solutions. They actually might produce the wrong solutions, where there are false positives, or there are things that are happening and so I think that’s something that we have to kind of be on the lookout. I just think there’s still going to be really good value in new sensors that are being developed whether we’re still on the cusp of getting value out of hyperspectral sensors, thermal sensors and others. I think what we’ve seen though is a commoditized version of you know image pixels and just taking imagery, right. And so I think, as new different sensors get put into orbit, there will be a high value on them initially and I think it over time they’ll be commoditized. It’ll be interesting to see where some of these companies, like Maxar and Planet go to in year four or five…in the next four or five years.

JF:  Yeah, that’s a great point. I started my career in remote sensing and worked on hyperspectral sensors. I can’t imagine trying to explain the value of viper spectral to the average person, right! I mean, when you’re talking one nanometre spectral spatial resolution, that’s got to blow people’s mind, right!

AD: Well, imagine putting that on a flock of satellites and for doing continual monitoring getting that amount of data and sifting through all the noise, right. It’s crazy!

JF: So, before we wrap up, I want to ask you about St. Louis and the Geo-Future initiative you got going. You and I have talked previously about what’s going on where I live, that’s just a tremendous compliment to your efforts there and just maybe explain a little bit about what you’re doing because it’s really taken off, at least from my perspective.

AD:  Yeah, no. Thanks for allowing me to chat a little bit about it. Yeah and St. Louis…we’ve done geospatial for many years but I would say, done it in the shadows a bit. And it wasn’t until the big government agency that’s here, NGA, decided, “hey, we’re gonna set up a new campus. We’re gonna set up our west headquarters here in St. Louis, but it was director Cardillo’s intentionality of saying, “we’ve got to do it differently this time. We’ve got to open our doors. We’ve got to start engaging with industry, with academia and with the future workforce to not only talk about what our mission is (and that’s the classified parts) but just the technology and the mission and what we have to do. But also, get new ideas that are coming in because we realize that industry is moving so much more faster than what we’re doing.” So that decision was kind of a catalyst for St. Louis to take more of a look internally and say, “wait, this is pretty interesting. NGA is setting up and investing in St. Louis but do we do more of this?” And so, a lot of people who don’t know geospatial but no economic development said, “hey, is there something around this? Are we doing more of this? Is this a competency that we can grow?” And come to find out, we were doing a lot more of it in agriculture and transportation logistics and others around some of the companies or capabilities we had around here. But now the question…and so we developed a Geo features roadmap and it’s really to light up the opportunity that we have around growing a workforce, growing innovation, and growing opportunity, and also with the intentionality of doing that with racial equity and inclusion in mind. And, especially the neighborhood that NGA is moving into, which has been challenged over many years and so it’s trying to do it the right way, and so anyway, there’s a lot of great opportunities and what we’re seeing. The great part is before NGA…they’re still constructing the building right now…we’re seeing a lot more companies that are coming in and saying, “Hey, I want to be a part of this” but hopefully over time we’re starting to see other organizations or other research institutes that want to plug in. Whether it be physically in St. Louis or virtually into some of the things that are going on here, to create and…going back to our earlier points here, as these companies that don’t know anything about geospatial are starting to do geospatial stuff and as we see it more pervasive in a lot of these as we start thinking about next generational things like; autonomy or flying vehicles and all those other things, right. That we can be a centre or a nucleus that helps in advance that here in St. Louis. So ,that’s our 20-year dream but we’re making some good first steps here.

JF: Yeah, I know the history goes back all the way to the defence mapping agency, I think they were originally started in St. Louis. What’s been the uptake by policymakers, politicians, mayors? Do they buy into it? I know like here, in Huntsville, our mayor’s like…he was really one of the original supporters. What’s the feeling in St. Louis?

AD: So, we’ve had…actually since we’ve launched our report, we’ve had two mayors and we’ve actually had three mayors since NGA selected, and the excitement from the local politicians has been continuous here. And I will say, from the state senators to the local congress men and women that have been here, they’re very excited. However, they are challenging us to say, “what is the future workforce look like? How do we make sure that we’re providing benefits for everybody and not just…and so, that’s the part where…we wrote that directly into our report, it was very intentional because we’ve got to do it a little bit differently! So, that’s the only thing that I think we’re getting continual emphasis on and by all of our policymakers is saying, “let’s be a great example of how to do this throughout the nation, right and to provide opportunity for those that have been disinvested in, in many years.” And so, we’ve got a great first step. There are great workforce programs that are already going on. Many of the Macs are basically published and said, “hey, we’re going to create our diverse workforce and opportunities in St. Louis. We want to build that there because there’s programming that we can connect to do that.” So, then on the another angle too is, we also have the universities. The St. Louis University is…next week they’re holding their Geo resolution conference, which is great and it offers a great opportunity for students to see and experience what is this industry that we keep hearing things about. Like we might be sitting in the political science department but I hear about geospatial. What is that, right? And so, provides opportunities for industry academia and researchers to come together and talk about it. So, we’ve had a lot of great success so far but it’s got to be way bigger than what it is today and figuring out how we grow it beyond just, NGA. Because if it’s just in NGA, they’ve been our foundation, but it’s got to be bigger than that and connecting in some of the other assets that we have here.

JF: Yeah, that’s fantastic Andy and thanks for providing the overview and so let’s leave it there. And again, congratulations on the success and best of luck down the road.

AD: Fantastic. Thank you, Joe.

JF: Thanks again for joining us on another On Point with Korem and don’t forget to subscribe where this podcast is posted. Whether that’s Apple podcast, Google podcast, Spotify or YouTube. And if you like today’s podcast, please leave a comment in the comment box. Join us again for another episode of On Point with Korem, where we’ll get On Point.