EP8 – The Evolution of Location Analytics in Business Intelligence Solutions

In this episode of On Point with Korem, I spoke with Dr. Sarah Battersby, a principal research scientist at Tableau, a Salesforce company and a former associate professor of geography at the University of South Carolina. Dr. Battersby spoke to the issue regarding the overlap of functionality between GIS and business intelligence and how much functionality should be developed within BI tools vs. how much is actually used and needed. She talked about how location analytics has become a true differentiator in the BI space. As a former University professor she advocated for the need to insert geospatial technology into curriculums designed for the business professionals.


Joe Francica: Well Sarah, again thanks a lot for doing this. I’m really looking forward to this because it’ll be one GIS geek to another, except I work for a GIS company and you work for a business intelligence company and there have been this collision of forces, I feel like in the universe, between BI and GIS. So, I wanted to start the conversation there about…certainly Tableau has built in a lot of functionality with respect to spatial analytics, so do you think at the very first, about whether this is a real competitive differentiator for Tableau versus other BI companies?

Sarah Battersby: I mean I think, location’s really just such a huge factor in business intelligence but also in self-service analytics as a whole and when I think about things like the mapping functionality in Tableau, our focus is so much on just finding the suite of solutions for the needs of our customers to make it easy for them, to focus on asking questions. Because people have questions. They don’t necessarily have a clear spatial question, it’s just “I need to get to the answer of something that’s important to me,” so instead of trying to make people wade through all of the complexities of spatial data processing, we’re just always thinking about how do we make spatial possible in an easy way, in a delightful way, so that you can do things at that speed of thought. And I think that’s part of what makes us a differentiator, is that real drive to make it easy and delightful and meet customers where they are?

JF: So, you just said something that that is always something I like to use which is, when I tell people about, “okay, you want to use location analytics, what you really want is to get at the answer” and you just said it. And I know that there are tremendous visualization capabilities within Tableau, I’ve played around with it. I’ve been to the Tableau Users Conferences, tremendous amount of energy into the visits that are created. How much awareness do you think, among the general user community of Tableau users, understands what they’re creating? Understands that they know that they’re creating this map visualization? That they’re thinking spatially with the questions that they’re asking of the technology? So, how much do you think that they’re just creating wonderful visualizations versus giving the answer to a specific business problem?

SB: I mean, I think it’s a little bit of both, really. I mean if people ask a question and they can get an answer, it’s like they’re super happy. So yeah, what they initially set out to do is, you’ll create this visualization that gets them from point A to point B. Tell me where I’m going. But I think people are also just trying to create things that are useful and make sense for their analytic flow. And that’s what I think we do well. Keep people in that analytic flow, help with good practices, just get you to getting the answer without having to wade too deeply into the intricacies of how that magic happened so much. You know, trust that the magic was right. But trust that your answer is what you need.

JF: So you bring up an interesting problem and it gets to this core of what I said at the beginning of this collision between these two users communities of GIS and BI, and that we understand what we’re doing and that’s true too of geospatial people, is that they’re in deep in the weeds in spatial functionality and so I think there is this trade-off of how much functionality you put into a BI product versus how much you put into a GIS product. And frankly in my mind, a lot of GIS products have almost exhausted the capabilities that they put in there, such that, I mean you and I have heard this bantered around about well users only use 20 percent of the functionality in any GIS product. So, I’m curious as to your vision for the product and how much more functionality can you put into like a BI product?

SB: I mean I think there’s always more room to grow in terms of functionality, but you mentioned that 20% sort of threshold…really I mean, I’m the sort of person that I’ve been working with geospatial technology long enough that sometimes I open tools just because I want to know what they do and I’ve probably used like 40% of the basic functionality that’s built in; like GIS and ArcGIS and other kind of standard geospatial type of packages. So when I think about that 20% measure, I think that we cover so much with that and so when you think about where you grow outside of that core set of functionality, you start getting into a little bit more of these edge cases, or for users that really want to push the boundaries or start having more specialized questions, and that’s where you start to couple that functionality with other tools for people to use, to keep them as much in the flow as they can. Because I’ve seen so much growth in the spatial environment, especially in the open source communities, so thinking about all of the Python libraries and packages that you can use with “R” and thinking about how do you provide the core functionality that users need but also allow for these real custom cases where people can put together super powerful workflows by doing things like; running a script in the middle of their analytics. You’ve got this pre-packaged script that does this super important thing for you. Tap into it! And then bring the result back in. I feel like in Tableau, being able to use “Tab PI” or being able to connect out to a script in R, or one of my personal favorite things, I run a lot of data out of spatial databases and you’re running sequel queries to do these things that I can’t do in a product that are not something that everybody really wants to do. Like the engineering isn’t worth it, when you can do it so easily just by having your data in the database. Or you’re running these extra scripts because it becomes more edge case at that point and I think that 20% number is really pretty spot on with what we use in the GIS.

JF: So, to that regard do you feel like the user community, the Tableau user community, is being pulled along with utilizing some of this advanced functionality, or maybe in reverse to answer that question; are your users pushing Tableau to integrate more functionality either with R or just to add that as standard functionality within the product?

SB: I mean, I think we get kind of both this push and pull. We spent a lot of time talking with our customers, the folks, the engineers and the product managers on our maps teams are always trying to get a good sense of what are the most powerful impactful features for our customers. So we’ve got this poll but there’s also a push of seeing what is it that people are doing that is really outside the box that would be a lot more delightful, if we could bring that into the product. So, it’s like we’re really keeping our eyes on all of that and trying to get the assessment of what helps people get to that answer faster because that is always the bottom line. It is; you have a question and you need to answer it. How do we make this a self-service analytics process where you can just answer your question!

JF: When I went to the Tableau User Conference a few years ago and we were actually getting in person with people, there was a tremendous enthusiasm for the product The Iron Viz, competition was astounding to me. I mean these people were just magicians at doing that. I guess it does go to the heart of the issue of this…again this trade-off between map visualization, charting visualization, do you think that exposing geospatial capabilities, this ability to visualize, also puts this in perspective what you can or cannot do with charting. And the trade-off that you can only see so much in spreadsheets but the map as a paradigm actually is quite useful, maybe more than people understood before, is that…am I off base with that or what do you think?

SB: No, I mean that’s something that super resonates with me. Like back in the day when I was teaching university classes, the thing that I tried to bring home to my students all the time was, the math is important! You’ve got spatial data, you need to look at the data, but the map is generally more powerful when you compare it with other visualizations to help people understand what’s going on with the data not just the where. And so thinking about how that map really pairs with the chart that helps you understand other attributes that drive the spatial distribution, or the second map at a different level of detail that helps you filter and mine through the data. I mean, to me, it’s really all about the interactivity at this point and that is such an exciting place where we are in visualization, now! Where it’s so much easier to create these visualizations that allow us to really explore the spatial, and the drivers, of the spatial.

JF: Yeah, and I think that what Tableau has done in particular, is kind of floated the boat a little bit more, for I would say the knowledge worker, much the same way that you know we look at Google Earth did for satellite technologies that you expose more people to what you can do and maybe you can speak to this recognition of location-based data as something extremely important to analytics but also just the recognition that there is this component within the larger data that’s captured within corporations and I’m not sure they understood that before, maybe they didn’t have the tools, before but now they do.

SB: Yeah, so I’m going to take this in a slightly different direction because I think there is this huge recognition of the power of spatial, but I think we’re also in this interesting time where spatial has become so ubiquitous that we almost don’t see it anymore. And you take out your phone and you want to know the answers; like, “tell me about the restaurants that are near here,” you aren’t necessarily thinking about the fact that this is seriously a spatial question. You’re thinking about routing, you’re thinking about combining that with other attributes or the restaurants, is it open? Is it the type of food that I want? You just have the question and I think that people have become so used to spatial technology being embedded in everything we do that they almost don’t see it anymore. So I think we’re almost at this strange crossroads of spatial being more powerful than it’s ever been before because we have so much more spatial data, we have so much more processing capability, we have so many more really interesting algorithms, and you know tools that are behind the scenes that are getting us amazing information and the ability to answer questions we wouldn’t have been able to answer before, or get proactive information. Like back when I was actually in an office, I remember I would get these periodic notices at like 4:30p.m.; “traffic is heavier going home than it is normally at this time.” I’m like, “Oh, my phone just told me this proactively.” Granted it was not smart enough to figure out that I rode the bus, so traffic was always going to be exactly the same on my bus every day. But it became like invisible and super powerful at the same time. And I think that that’s just the crazy place where we are, where the value of spatial is so much higher than I think it’s ever been, because we’re doing so much of this really neat work behind the scenes, but it’s also so much more invisible than it’s ever been.

JF: So, what you’re essentially saying is, there’s an expectation that there’s going to be location data clarifying, modifying, the eventual answer that you want to get. That there’s an assumption that there’s a spatial component to every answer.

SB: I think there’s an assumption, but it’s also a hidden assumption. Yeah, it gets back to that; “I just have a question, I want to know the answer,” “who should I go visit today, I’ve got an extra half hour in my day,” “what’s the best customer to go visit,”  that sort of thing. There’s a lot that goes into answering that but you know I want to just say; “I’ve got half an hour” and the system says, “oh, go visit Joe, he’s nice. He’s a good person to talk to!

JF: So, let me ask you a different question. Going back to your days at university, I think you…and correct me if I’m wrong, but you taught in the geography department?

SB: Yeah, correct.

JF: What’s your take on bringing more understanding about geospatial technology into business schools? I’ve tried…I’m not sure I’ve succeeded in exposing that….And their answer I get is; “there’s no room in my curriculum.” I think that’s got to change and what’s your perspective?

SB: I mean, you’re opening a can of worms. I mean, my perspective is that spatial really belongs in most disciplines. I mean, whether you think about the cores of spatial thinking and the overlap with how you think about math and chemistry and sciences that aren’t necessarily geospatial, but have a lot of; how is organization important for understanding, molecular structures or whatever? You think about disciplines like business specifically, since you brought that up, that is somewhere where really is fundamental in thinking about how you answer questions. Understanding geographic relationships. How cities are structured? How distribution takes place. Anything with supply chain. Spatial is critical in those. Whether you’re talking about the last mile or getting things into the warehouse so that it can be there at the right time, just in time processes. I am never going to say that spatial should not be part of the curriculum. But having worked at a university, I full well know the challenges of getting everything into a curriculum that is required to be in the curriculum, and I think that’s going to be a fight that we’re always having. There’s also a little bit of siloing to deal with because a lot of the people who are teaching classes have never had formal training in particular areas. And you say, “You know this area really well, now I want you to do these three other things. Learn them well enough to really bring them in to your class.” I think that at least at the university level, we can be doing a lot more to think about how we can do cross-disciplinary education and really allow people to shine in their areas of expertise, but this is totally another can of worms that maybe we’ll have to open later.

JF: Yeah, because when I was in business school, I had one professor who knew how to spell GIS, and it was in real estate finance. And I thought if there’s any place that you ought to introduce this capability would be, certainly in real estate. But there just didn’t seem to be a place. There’s no core course that would say, “Well now we’re going to talk about retail and location.” To me, I kind of divide the universe into two buckets; the traditional users of geospatial technology and the geographically challenged. Right, so the people in retail and banking and insurance, although today and again I’m curious as to your perspective, is those companies are really utilizing the technology very deeply. The map is useful but to them, if they’re doing geocoding and analytics, that way, well they know the power of location and they’re underwriting policies and they’re making millions of dollars of commitment to buy real estate, etc. so they get it but they don’t want to talk about it always. So, I don’t know if you’ve had the same experiences but it’s deeply embedded, I think more than we suspect

SB: Yeah, I agree. I think it is deeply embedded. And I want to go back to something we were talking about a minute ago because I think it’s related and that they’re really kind of two camps. You said, you had the geospatially challenged and the kind of power users. But I think that there’s also the people who can use spatial data and people who really understand the fundamentals of spatial data, and that is probably part of what plays into the university challenges, but I think it also plays into the challenges with working with any spatial technology. There are there a lot of oddities with how spatial data works, especially when you think about map projections and what happens with those translations and distortions, but also just generalization, and data quality, and what is the data source and fitness for use. There’s a lot that goes on behind the scenes that I think is really complicated for people. This is probably why it’s a little more challenging with university curriculums as well, because there’s the teaching; how to do something versus the teaching how to think in a spatial way. And that’s always going to be a bit of a challenge, especially in an environment where I just want to get to the answer.  Yeah, self-service analytics, we want to help you get to that answer but sometimes there are some road bumps and understanding those bumps along the way, help you get to a better answer but they’re really frustrating when you encounter them.

JF: Yeah, I can see that. I helped develop a course for the folks at Penn State, there were three of us who were developing a course, and the course was called Location Intelligence for Business, but it was taught in the geography department. And there just wasn’t the space to go into the business school, but clearly students who took the course really liked it, because they never heard that perspective before. They never put the two disciplines together; between the business side and the location side, and I think if more students saw that, even more professors saw that, they’d go “Maybe we ought to have an interdisciplinary course like that” as an example.

SB: Yeah, I mean, that really unlocks the power so well, because you get the best of all the worlds. You get to dig into what’s happening with the data but you also get this very tailored, “here are the scenarios” where it’s really important to you, and you can see that real world. How it solves a problem that would be really hard to do without digging into the spatial data.

JF: So, one last question before I let you go. Is given… you have great deal of influence and direction of the product at Tableau, where is this going? Not just in Tableau but in the BI space? Where is that next level that we have to take users or again, as we just talked, where do we have to educate more knowledge workers in that? So, I’m curious with what you’re feeling.

SB: I mean, I think the next level is just thinking about; what is the current functionality and how do we make it easier and more delightful? What are the functions that unlock new questions that maybe you didn’t even know that was your question, but once you have that capability, you can do something powerful with it! And then, for me, from a personal perspective, because I’m on a research team and my background is really all about how people think about spatial data is thinking about; how do you expose to people when there might be problems that come up? Or, how do you help people understand those challenges? That’s a real simple problem that comes up quite a bit, that I see is; ”I have point data, I want to know what census tracked all of my points fall in. But if you have different levels of resolution for your data, the point may fall just outside that line. Or you’ve got county level data from the census and it’s generalized to one to 20 million instead of one to five hundred thousand, but I know that point is in such and such county, how come you’re lying to me?” I think that we have a lot of room to help people understand how the data works and to feel more confident about the results that they’re getting. To me, that’s the biggest hurdle to overcome. We can build in all the functionality we want in any product but until people really understand what’s happened with their data… yeah, there’s going to be kind of a trust and a confidence issue. So, that’s something that I personally try and take on.

JF: Well, I think you just opened another can of worms that will have to have another conversation, but really, thank you so much for your perspective. Because I think it is a different way to think of things, because you’re in a community that you’re trying to educate more. So really, really appreciate your time and we’ll have to do this again.

SB: Yeah, thanks Joe, I’m looking forward to talking some more. Maybe we can actually see each other in person and do this.

JF: Wow! Yeah, that would be great! Thanks.