The Center for Data Innovation spoke to Ondřej Tomas, co-founder of CleverMaps, a Prague-based data visualization company. Tomas talked about the complexity of managing land and infrastructure in farming and construction, and how the visualization of data in maps can simplify the task of complying with regulations and identifying problems.
Nick Wallace: CleverMaps maps out companies’ business processes. Can you tell us a little more about the activities you apply this kind of visualization to?
Ondřej Tomas: When we started CleverMaps, we didn’t really know what areas we would focus on. We just started organically, and over time we developed solutions for various types of customers. One of the first groups that we worked for, and that we had solutions for, was agriculture. That includes both farmers running agricultural businesses as well as the landowners who lease the land to the farmers. The second area is large infrastructural projects and construction, like roads and pipelines. That also grew organically—we’d worked with those companies before. We also work with utility companies, retail banks, insurance, companies, retail outlets, and e-commerce firms.
We started about four years ago. Our goal was easy-to-use maps for business users. That’s why we called it CleverMaps, because the maps were supposed to be clever, not complicated. We wanted simple, straightforward clever solutions for end users.
Before this I worked for a mapping company where we gathered all kinds of location data, including aerial photography and so on, and earlier I worked for IT companies. I always had a feeling that there was something lacking—I never really found an easy-to-use mapping tool. For example, when I was working for an insurance company, we were dealing a lot with locations, and all that existed was geographic information systems (GIS). We really struggled with that in our business analytics department. So I felt that this was a space for improvement.
In agriculture we have a solution called CleverFarm. On one side, it works on leasing and renting the land effectively for farmers and landowners. A typical Czech farmer doesn’t really own the land he uses. He owns parts of it, but 80 percent of the land he’s working on is leased. The Czech cadastre is very fragmented. When the communists came to power in 1948, they collectivized the land into huge state agricultural companies. After 1989, some of this land was returned to the original owners in restitutions, and some of it was privatized and sold. If you look at it on a map, you can see that there are tons and tons of tiny pieces of separately-owned land that you basically have to get together in order to have a field on which you can work. The average Czech farmer has close to 1,000 lease contracts. That’s a massive amount of bureaucracy: we automate that. We basically infuse the cadastral data into the map, and then automate the process of leasing, renting, buying and selling the land in the platform.
The other thing we do is closer to the agriculture itself: we map out “agra-evidence,” what the farmer is doing on the fields. There is other software out there that does that, but what we do is check cross-compliance for subsidies and regulations. Our platform insures you aren’t breaking subsidy rules, and that you aren’t breaking environmental protection rules. For example, if you’re fertilizing land, there can be only certain concentrations of the fertilizer, or if there is water, you have to have buffers around the water—sometimes it’s one meter, sometimes it’s three meters, sometimes it’s fifty meters, depending on the type of fertilizer you’re using. You need to work with information about altitude and the terrain: if a slope is too steep, you cannot run certain crops, because of soil erosion. These rules require cross compliance: they complicate one another, and there are so many of them that it is impossible for a farmer to keep it all in his head. Usually they do this on paper today, but they make huge numbers of mistakes that can cost them their subsidies, or get them fined.
As you can imagine, it needs to be super simple for the farmers to use. These guys don’t care about technology, they don’t want to waste time on a computer or an app. You could do this with GIS, but nobody would use it because it would be too complicated.
Our work in construction addresses very similar problems to our agricultural work. Say you’re building a new road or a pipeline: you have to buy the land on which the highway will be constructed. We analyze the ownership of the land of the planned construction, and we identify potential planning problems. Sometimes this allows us to say that whereas it might be impossible to implement the original plan for ten or twenty years, a subtle alteration might enable you to do it in under four years.
Wallace: Can you give an example of something unexpected that you’ve found through this kind of analysis?
Tomas: This happens all the time. We were recently doing a road management project for one of the counties in the Czech Republic. That was actually not about building a new road, but about analyzing the existing road network in the county to determine how much of it carries problems that need to be solved, in terms of land ownership or rights of use. The number was over 30 percent. It shows how big the problem with land ownership is in the Czech Republic, and what was done when the roads were built: 30 percent of these roads were built without permission, on pieces of land where they should not have been built. They now have to go to the owners and buy the land. We encountered very similar problems in our work with energy suppliers. If those cases went to court, it would be extremely expensive, so it’s much better to deal with the problem pro-actively by using our platform to identify and communicate with the landowners.
Most of the surprising examples come from our CleverAnalytics location intelligence solution, where we work for banks, insurance, e-commerce and so-on, where we work with lots of interesting stuff. Very often our clients ask questions like, “we are successful right across the country, except in this one city—why?” We had a bank with that problem; they also had a city where they were very successful, and didn’t understand why. In either case, you need to understand the situation in order to manage it, and that’s where we come in.
Wallace: How is this different from GIS? What other data are you drawing from, and why?
Tomas: GIS solutions are built for GIS experts and analysts. You really have to have a person who knows how to work with GIS. He has to understand the location data, he has to understand how to build layers, he has to understand how to run queries in GIS, and so on and so on.
For example, energy suppliers usually have a large GIS department of their own, with say 20 to 30 people, who are running their own analysis based on GIS. All of these people are experts, and if they finally come to a conclusion, they build a certain picture which maybe they later on can send to management in the company.
But we’re building software for the end user, letting them view problem directly, but without requiring them to play with building or combining layers, because this is not something a normal user would understand or know how to do. The average person doesn’t know how to work with datasets, how to normalize them, which ones they can combine. We’re building solutions that utilize the map, but all these complex tasks are done in the background of the software. They’re prepackaged solutions doing very specific things, so you know what the software is doing, but you understand the outcome. It’s business end-user targeted software, not a tool for experts.
In our development phases, our people do work with some GIS solutions. Our software has a location database with the specific extensions and so on, which can run the queries, but all those processes are already pre-packaged and automated. The user just gets the result. With GIS you have to know how to cook the meal, we give you the ready-made hamburger.
When we started, we were working with over 100 datasets, we were testing and trying everything out, and found that we kept coming back to five basic ones, and built our model on those five, and then we add extra data sets when something is needed. The five datasets are census and demographic data, roads infrastructure and traffic flows, public transport data—because that gives you very precise information about the flows and concentrations of people—data about buildings, which we combine with the demographic data, and points of interest with relatively detailed information about them—for example, if it’s a hospital, how big is the hospital, and how many patients does it have capacity for?
Wallace: What benefits do companies get out of visualizing their activities like this?
Tomas: In terms of the location intelligence software, it’s the context. It helps you to understand why a particular thing is happening. If you are planning to roll-out new automatic teller machines (ATMs), and you’re getting into new places, you want to understand why the existing network is the way it is, why it succeeds in one city and fails in another one. It’s all about context.
The other benefits are mainly about reducing complexity, especially in farming. It saves a lot of time by automating basic bureaucratic tasks in large volumes, and it avoids unnecessary mistakes and errors that can cost a lot of money in fines and lost subsidies. We were able to reduce construction times in some infrastructure projects by as much as 30 percent, by mapping out the relevant information and automating processes.
Wallace: You clearly have a strong focus on the agricultural sector. How do you expect the Internet of Things and “precision agriculture” to impact your business?
Tomas: I think it will have a massive impact. The current adoption of precision farming in the Czech Republic is basically zero. Almost nobody is using it. But lots of farmers already have machinery that is equipped for it—they are simply not using it, because it’s just too difficult for them. Or sometimes it’s too expensive: when they order a new tractor, they sometimes leave those options out because they don’t want to pay for them. It’s like buying a car and deciding whether to pay extra for the navigation tool, or to use your iPhone with Google Maps. Someone will have to package precision farming into a working solution for farmers that’s easy to use.
A lot of the precision farming stuff is quite fancy, but simple things can massively improve the farmer’s business. For example, a sensor for temperature and humidity is very simple, and if you place it in your silo you can can get an alert about rising humidity that could cause your hay to rot. If you are not fast enough, it will be destroyed, but with this solution you could find out very easily. But the farmer is not an IT person, he will not be working with the sensors himself or monitoring it on his computer, he needs to consume the information in a very simple way—maybe through his iPhone. That is how we see the future for our solution: we need to package it into a simple solution for farmers, and then I think it’s going to have a major impact.