Mapping Personalised Workplace Risk

May 15, 2019 21 min read

Mapping workplace risk using personal location and geofences

  LISTEN HERE OR WHEREVER YOU GET YOUR PODCASTS

 

Workplace risk and danger are personal, spatial and time-based. In this episode, you will learn how one company is solving for this using personalized notifications based on location. A geospatial safety bubble that follows you around and alerts you when you are at risk. 

 

 

           

 

PUSH THE PLAY BUTTON TO LISTEN TO THE INTERVIEW OR SUBSCRIBE TO "THE MAPSCAPING PODCAST" WHEREVER YOU GET YOUR PODCASTS.

Clint:
But yeah, as a general rule, you want to have a dynamic geo-fence around the user which follows them. So we call it a safety bubble for purposes of marketing. And that safety bubble, when that interacts with a hazard or risk, that's when it triggers alerts for the worker.

Daniel:
Hi, and welcome to another episode of the MapScaping podcast. My name is Daniel, and this is a podcast for the geospatial community. Today I'm talking to Clint, and Clint comes to us from a company called SaferMe. And at SaferMe, they've devised a system where they can deliver personalized risk assessment and risk alerts to users based on their location. It's a really interesting idea, and I hope you enjoy the interview.

Daniel:
Welcome Clint. Thank you so much for taking the time to do this interview with me today, I really appreciate it. Now you've started a company, or you're the founder of a company called Safer Me. Can you tell us a little bit about it?

Clint:
Yeah. Yeah, no worries Daniel, thanks for having me. Safer Me uses data to protect people at work. So 820000 injuries happen at work every day, and we try to stop people being injured by warning them before something happens. Might well be a warning event on an asbestos location. It might be a dangerous dog. Perhaps it's a lightning strike. Some sort of specific piece of data that's really quite relevant to the worker to keep them safe.

Daniel:
I think that the key thing that the people listening to this podcast are interested in is geospatial. And the thing that really interests me about what you're doing is that SaferMe is giving a risk assessment based on the location of the worker.

Clint:
Yeah, yeah. I mean, we try to give the worker something incredibly relevant to them, which means location. So we have to track where the worker is relative to hazards and risks, and then assess how dangerous the particular hazard or risk is at the time the worker's there. And that's primarily what SaferMe does. Obviously, in order to do that, we have to allow the worker to record all of their risks and hazards that they see as part of their standard health and safety and OSHA process. Depending on which country you're in, they have different acronyms for the safety processes that exist in that country. So that's our role, yeah.

Daniel:
So it sounds like the system needs a lot of data to work. So it's one thing that they're recording their own data. Let's say they're out on an oil field or something like that, see a hazard, record it, and that's great. That's there for everybody else. But there must be some sort of base data that you also need in the system to be able to give an accurate risk assessment depending on location.

Clint:
Yeah, exactly. And there's certain data that's better than others. And whether that's via coverage or that is the most relevant. So for example, if we take lightning as a dat set, that's actually one of the better data sets you can get hold of. It's global by nature, all measured in real time. Well, 98% of lightning's measured with relative accuracy in real time. So we can take those lightning strike locations and then try and calculate the risk of the next lightning strike for the worker. And particular workers care a lot about it. For me and you, lightning's kind of an interesting thing. It's spectacular, but it's not something that's likely to kill us in our day to day job. Whereas, if you're a worker that's on a metal object, a crane, perhaps servicing high wires, oil and gas, these people can be killed by lightning. There's 20000 lightning deaths per year. So our job is to try and compute where the next lightning strike is likely to be, and to obviously put some error bars around that and then warn the person before that happens to them.

Daniel:
That sounds like an amazing data set to look at, and a very dynamic data set. So I'm thinking this is something that's updated constantly, it's not a static system. It's not enter data once and forget about it. You must have some kind of rules around the temporal data of nature.

Clint:
Yeah, I mean there's so much data. So we specifically are interested in time bound, location bound data that's relevant for safety. And then even within that small subset is really four distinct types of data. And the way we think about it is it's just a simple two by two matrix of data types that are relevant for safety. I could explain it for you. I'd have to draw it out for you. But if you imagine the two by two matrix that's one column, you have machine or internet of things data, and the next column you have human data. And then your two rows would be open, or freely available data, and then there's private data. And SaferMe has different methods to collect and share those four types, but within the data types. So to give you an example, lightning would fit within internet of things, proprietary, or private data. Something like fire or emergency information is typically open in internet of things or machine based. And then you have crowd sourced human data where the worker just pushes the button and quickly takes a photo of whatever the problem is.

Clint:
So there's various date types, and all of it is time bound and location bound, and it's specifically relevant to different types of workers. So in its own right, it's kind of a niche, super geeky area of data.

Daniel:
Yeah. I think any time you trying to work with risk data, like risk is so personal as well. I'm assuming though, the organization, whoever is buying the system, is going to set it up for their workers, has some sort of standards around risk. "These are dangerous, these are less dangerous, these present no risk to us." So I could imagine that if they come with their own data for example, that that data is already categorized. But I think in general, risk is quite an interesting data set to work with.

Clint:
Yeah, and then even within that. So there's always the danger itself, and then there's the risk. And they're actually subtly different data sets. So if you take the lightning example, there's the lightning strike, so the point based data. And then from that strike, we calculate the risk of the lightning. And if we're dealing with say a totally different data corner stone, something like mine shafts or sink holes, which we obviously sell to clients or help clients avoid as well ... There's a mine shaft, and say it's a 200 meter mine shaft straight down. There are certain people who care a lot about not falling into that hole, that shaft. And the risk of the shaft is actually slightly larger than the hole itself. And there's an extreme example at the other end. If you have something like a fire, for example, the fire has a location. And then there's the risk of the fire, which is typically down wind of the fire. So there's always the risk of the thing and then the thing itself. And you have to have both and then track where the worker goes and give them an indication of the risk they're experiencing relative to the job they're about to do.

Daniel:
So it sounds like we're dealing with a lot of sort of geo-fencing here. But I think in terms of lightning or fire, for example, it would have to be a dynamic geo-fence, would it not? Like these things move, they're not static.

Clint:
Of course. Lightning's a nice case, but fire tends to be broken down by regio. So some regions or provinces are quite good at releasing realtime fire data. Some regions have 20 minute delays and point based data which is not as good. So it really depends. Different data has different quality globally, but yeah. As a general rule, you want to have a dynamic geo-fence around the use which follows them. So we call it a safety bubble for purposes of marketing. And that safety bubble, when it interacts with a hazard risk, that's when it triggers alerts for the worker.

Daniel:
What could that alert look like? I'm assuming this is something like an app that's running on my phone, so it's a portable device.

Clint:
Yeah, exactly. So you're pushing alerts to the worker over the best method you can, so typically an app. So the client will have their own branded safety basically, that's what we deliver. But it might be pushed over some other interface to the worker as well. But we have to have their location as a starting point.

Daniel:
I was thinking about this, we talked about this the other day when I pre-interviewed you, and I was thinking about leveraging infrastructure as well. Because infrastructure could present a risk. And specifically, what I was thinking about is the open street map data set, and I was thinking about roads and traffic and that kind of thing. I'm guessing here, but when you're talking about sink holes and mines shafts, we're talking about people that are out in the field somewhere. Would that sort of infrastructure be relevant to risk?

Clint:
Yeah. I mean yes and no. Some clients upload their own shape files to give the worker more context. So if you think of it from the worker's point of view, they're staring down at a map with their location. They can see risks around them, they can click on the risk and understand what that risk is. And then behind the scenes, we're obviously measuring that experience and providing proof for the company that that worker has been made aware of a risk as they go about their job. So that takes care of a lot of obligations for the business automatically. But the actual infrastructure itself, we try to stay away from shape files, which we do have them on the system for context. We do try and stay away from asset tracking because that's a whole different area, a whole separate niche which is complicated in its own right. Although, there is some overlap, so I hope I haven't confused you. Or not confused you, you not what you're talking about. But I haven't been confusing to an outsider I guess.

Clint:
But things like heavy machinery, we can see that happen. So tracking where heavy machinery is relative to the user, and then having some interaction between the geo-fences of each, the machine and the user, and giving both sides alerts when needed. That's kind of inevitable, but not quite where the market is yet.

Daniel:
Do you think that kind of thing will come with time when we start moving into the internet of things where every sort of device communicates with every other device, where that will just be an expected thing that, "Of course. That heavy earth mover is backing. They can't see me, but I have a device in my pocket, so it's sending out push notifications to everything within a certain radius."?

Clint:
I mean, if you zoom right out, inevitably the virtual protects people in the real world. But if you zoom in to where we are today, it's pretty easy as product people like us to forget that workers in the field, your blue collar engineer, welder, plumber, roofer, scaffolder, these people, they're only just getting phones. The phones that are being rolled out are being rolled out last year, this year, and next year. So we've had smart phones for ... I mean, I'm generalizing here. But I know I've had a smart phone for five odd years, maybe seven. But the workforce is only just getting these. So the products that we need to deliver, while they can be advanced, just doing simple things well is actually really quite valuable from a sales perspective. So that means simply and easily recording risks and hazards from the worker's point of view just on a phone. Although we can experiment with all sorts of other technology, I think being specific and useful is important.

Daniel:
Absolutely. And I think that's a really interesting observation, that it's not necessarily the technology that's holding us back here. It's more of a cultural barrier. People need to get used to these applications and get used to using them. And when we do that, that will open up more doors as well. But maybe it's important not to push the envelope right at the start in terms of introducing too much, and maybe even over-engineering something, if you know what I mean.

Clint:
Absolutely. So in the beginning, when we do a roll out for example, if you think about a rollout as a population of people to roll out to, there's a ticket option curved within say a population of 5000 users. Those first 14% of early adopters, they're actually quite keen on the new technology, but you've still got laggards within your business environment, that you have to get them used to using the technology. So we'll always present it in the beginning as being an easier alternative, and not force use of the phone. Because primarily, our competition's paper. So you're talking about people in the field with clipboards, trying to write down where dangerous dogs are, and asbestos, and all sorts of hazards. And that whole process needs to be automated obviously. And the future of the process of they have isn't a better paper form. It's not a digital version of what they're doing. It's actually just automating it all the way with data itself. And that's SaferMe's role.

Daniel:
Again, I think that's a really good observation. And I think being in the tech industry, in the geospatial industry, it's moving fast and we can do so much. But it's important to remember to take the user with us, to remember the user journey instead of charging off out there into the sunset. We can do this, and internet of things, and blockchain, and everything else, and big data, and use all these acronyms, and frighten these people away. Because without them, without their adoption, then we can't do the exciting things and solve the problem either. So it's really important to remember that user journey.

Daniel:
I just wanted to say one more thing here. I really like the idea that this is solving two problems, or two problems that I see anyway. You're keeping people safer. You're providing that sort of very personalized service, a personalized safety message to someone for them saying, "This is relevant to you. Not relevant to everyone, this is relevant to you." And at the same time, you're fulfilling a lot of documentation, you're providing that documentation, which is becoming increasingly important. And this whole thing is sort of automated. "We did send that message to Dave. He was told before he went into that are. We had an obligation to warn him and we've done that and we can prove it."

Clint:
Exactly. And that's the type of thing you can only really do spatially. So we're quite unusual. I mean, health and safety and OSHA are relatively backwater areas. I mean no offense to the people working in those ares building software, but those areas of software are even more backwards than normal areas. And I think anyone on this podcast is really into location and geo data and geospatial, but safety is really not. And in the beginning we really had to educate people around why delivering location specific information to every single person relevant to them is important. The situational awareness of the worker essentially, so that they can make informed and smart decisions. That was a real education process. But nowadays, people are much more used to concepts like Google Maps. All the things that this audience has probably been seeing for 10 or 15 years, the rest of the public's catching up with.

Daniel:
Yeah. And it's really interesting that you say that there's been a whole lot of education around this. Because in some ways, we're very much used to having a personalized experience. When you use Google, for example, Google personalizes the things we see depending on our search queries, our search history, and our location. And so I guess the thing is with Google perhaps and other systems, is that it's hidden away. They haven't run out and said, "Hey, we're personalizing based on your location." But here, you're saying that. And I wonder if that's the true barrier, is people think, "Oh. Location. What's going on here?"

Clint:
Yeah. I mean, it is a new thing for the safety sphere, but it kind of can't be ... The problem that exists at the moment so large that it has to be really solved by delivering ultra relevant data to the person at the right time, it just so happens that there's so much data to choose from that we have to be relatively careful what we deliver the person and exactly how, because some people don't want to accept as many alerts as others. But we think the problem can be solved, so we're just going about it step by step really.

Daniel:
I think you talk about ultra relevant data, and you're right. It is ultra relevant. That's probably the only approach that's going to work in this situation. But it also raises a few questions around data privacy for example. I don't think anyone out there really wants to be tracked, or at least they don't want to be told that they're being tracked. How do you get around some of those issues around data privacy, maybe infringing on workers' privacy?

Clint:
Yeah, there's two major areas there. One is that we don't tell the business where the worker is, but the worker knows that we know where they are. And so we're sort of an intermediary between the business itself and the worker. What we found is that workers don't want to communicate at all times where they are to the company, and the company is in charge of keeping the workers safe. So it's the company that pays us essentially, but even so, we don't tell the company where the worker is. And then secondly, there are some privacy issues around the hazards themselves, and our position on that is that hazards don't have privacy rights. People do.

Clint:
So what we find is with all these workers basically scurrying around and recording and taking photos of these hazards, a lot of businesses in the beginning don't necessarily want to share their hazards. They see it as information that might put them at risk. But it's actually the other way around. There's a fundamental need for these businesses to share their hazards and risks. Because say for example if the public or another worker gets injured by the hazards the business already knows about. It puts them at great legal risk. So we do try and encourage customers to share their risks, but they don't have to. So that's a little bit of a transition that tends to happen with each customer.

Daniel:
Okay. So it sounds like what you're trying to create there is like a crowd sourced data set of risk between these different organizations. Is that a correct way of sort of thinking of it.

Clint:
Exactly. The problem can't really be solved without data sharing. It just so happens that some parts of the safety process are incredibly private to a business, and some parts can be shared quite easily. And certain hazard types are much easier to share. So a good example, something that's right on the edge actually is something like dangerous dogs. Something that's easy for a business to share is something like a mine shaft. So where it doesn't have a people component to the data, and so they're quite comfortable sharing that from business to another across the system. And then there's other data types where we have to strip out various meta data pieces to allow it to be shared between one company or another. When you're dealing with these massive businesses, they have offshoots and subsidiaries, and all sorts of workers within their own business that aren't aware of all the risks the business itself knows about. And that's actually really dangerous.

Clint:
So what we find is, these big businesses already have internet of things programs that are already collecting data. They already know about every asbestos location in the United States for example. And then they're tracking all this and adding to it in a control room type setting. And then the workers are out in the field and someone gets exposed to this risk and gets injured. Inevitably there's a discovery process after the injury, there's a legal process. And then the business gets sued and has to settle. And it's not that the business didn't want to tell the worker, their own person, about the problem. It's just they weren't equipped to do it. So whatever the data is, it's our job to try to share it with the right person at the right time, and that's our role really.

Daniel:
Yeah. And again, we're getting back to this idea, this personalized experience. Like what is relevant for you in this location? And also, again that idea of solving two problems at once. Like, providing the personalized experience, risk mitigation, and the documentation.

Clint:
Exactly. The admin load is so high to solve safety if you're doing it via the traditional method that it actually just can't be done properly. So if you think about the 820000 people who are injured every day, injuries and accidents, we think of it as inevitable, but 50 years from know people are not going to die on the roads. People are going to be very unlikely to be injured at work. And the real question is, how do we go about solving it? And it's got to come through a shared data set and specific alerting for each person. And at the moment, we start on the margins where we have good data and we have the technology to deliver thee right alert at the right time, but in the future we'll obviously have better tracking, better data, better systems delivering alerts. So it can be solved, it's just going to take some time.

Daniel:
Yeah. Immediately I start thinking about an open street map of risk data that everyone updates across the world and it's crowd services, and giving us this amazing sort of base layer of information to work from. So we've talked a little bit about the problems that you're solving and how you're doing it and the data that's involved and why you're doing it that way. Can you give us and idea of what kind of organizations or maybe industries are using this?

Clint:
Yeah. So typically utility and infrastructure managers. We do get pulled into mining and other sectors as well, but utilities and infrastructure tend to be the best, mainly because workers are distributed. They're often large teams of people in offline areas. And those particular types of businesses, we're very good at servicing in them. In the long run ... When I say long run, five to ten years from now when location tracking is much more accurate, indoors, outdoor, multi floor, then we can really start targeting pretty much everyone. But in the short run, we have to be thoughtful around who we target and who we sell to. So utilities and infrastructure. For example, Vodafone, which is a large telecommunications company, or Veolia, which is a order infrastructure manager out of Paris. So these companies are typically large global corporates.

Clint:
And touching on, I guess, the mention of an open street map of hazards and risks, what we've found is that individuals don't typically search to be safer. So you could create such a thing and we'd be for it because we'd be able to use that data to keep workers safe. But an individual themselves in unlikely, in our view, to search for safety. What they would be relatively happy to do if it's relevant, is receive or actively be given alerts that are relevant to them. The technology's not quite there to do that yet. But businesses will pay to keep their staff safe, and that's why we do what we do. Our role with how we structure the business is to try to generate as much money or revenue we can based off the best data we can get, so that we can leverage that to get more data into the system to solve the problem. So we try to build a repeatable business model that uses data and solves a real issue.

Daniel:
Yeah, that sounds like the bases for a sustainable business. I wanted to come back to a point that you made earlier, and that was, you talked about in the future there will be much better tracking available. And you talked about inside and outside, and there's definitely an issue there with tracking on inside, inside buildings. And the issue is that accurate positioning of data or of a person's location relative to that data. I've been lucky enough to talk to a few companies that are working in that field there, of positioning inside. And it was really interesting to note that they ... I asked them all, "How can you describe this? What's it going to look like in the future?" And the response was, "It's going to look like all the options you have in terms of positioning on the outside. That's what we're going to have on the inside. So all those options of positioning, locating things, and routing of things." Yeah, and so I mean, I think the future is really bright for the geospatial industry in general, and in what you're doing. And I think when you can seamlessly cross over from the outside to the inside and expect the same level of coverage and accuracy, it's going to be a really interesting time.

Clint:
Yeah. I don't know what to say to that other than yes. I mean, it's pretty exciting what's coming. We sort of stick to our knitting. So we get asked sometimes to try and deliver solutions that would work indoors, but that's just a whole different area. And there are so many businesses already focused on it that we will try and stay away from it until someone else solves it, and then come in and help people be safe in those environments as well. So yeah, at the moment we just stick to our repeatable business model we've already got really.

Daniel:
Yeah. Just before we wind up the conversation here, I've just got a few more questions. And these questions all sort of look out into the future. One of them is about the internet of things. How do you think that's going to change things in terms of what you're doing today? Obviously we're going to have more data available, but is more data automatically going to mean a better service or a safer world?

Clint:
Interesting. Yes, but there are some things that need to be built alongside it. A few realizations are coming. I mean, I think, I mentioned just before that these big companies, they have specific internet of things programs. So to put sensors out into their network of infrastructure and gather more data. And sometimes they get safe data, and when that's the case, they're just starting to realize that that's putting them at legal risk if they don't share that in realtime, because there's some latency basically between when a hazard is recorded and when a worker is warned. And if there's an injury in that time, that's a bad thing for the business. They can be found to have known about something they didn't want their worker to.

Clint:
So that's coming, which we can see because that's the space we're in. When it comes to the amount of data that's there, the problem isn't so much the data. The problem is the interface between the data and the user. So people don't ... When I'm walking down the street, or when you're walking down the street, you don't constantly look to see if you're at risk of being struck by lightning. Or if you do, you're an interesting person. But the systems that we build should be doing that on our behalf. So the analogy is not so much to search through the data that's there, like a Google type of example. The analogy is, how do we build an agent or agents that serve each person and deliver that person just what they need at the right time? And a lot of that work is already happening in the marketing field, but there doesn't appear to be too much of that going on from what I can tell in our space, except I guess what we're trying to build ourselves.

Daniel:
That's interesting. I think that marketing often leads the way in many of these sort of opening or emerging technologies. And it's interesting as well, because you'd think insurance companies would have caught on to this and say, "Hey, this is a really good idea," and start to insist that people spend more resources in these areas and build these applications that you're talking about, these interfaces.

Clint:
Yeah. I often wonder if insurers five years or hopefully faster from flipping around ... I mean, at the moment insurers are seen as these huge, big, scary businesses that have all this data on all these people. And I often wonder if they'll flip around one day and be kind of protectors of us all, because their interests are our interests. And I do wonder if they should really put a lot of effort into trying to stop the people they're insuring from getting fat or getting sick or getting injured. And I think it's kind of inevitable really. Someone will do it.

Daniel:
My last question, or one of the last questions is, Google wants to make all the data in the world searchable. Could you see someone like Google moving into this space? I mean, they already live in our pockets in the form of Android. If they could index all the data in the world, could you imagine them being able to deliver these sort of personalized notifications to us, in having this just sort a passive thing in our pockets that sort of alerted us as we moved through the world?

Clint:
Perhaps. There's a lot of different competitive verticals that are kind of being touched on there. So alerting, from our perspective I'm not worried about them as a competitor because our main function within the business, we're a B to B product based [inaudible 00:30:38] for large businesses. But eventually servicing small businesses, and then eventually the public. But ... Well, the data will be. But someone like Google, they're so big. What a great company, right? They've built something amazing based on the ultimate librarian experience years and years ago. And now they can pick and do whatever they want. But I don't know if they'd come into our sphere. There could be any number of competitors that try to.

Daniel:
Hey Clint, I really want to thank you for taking the time to do this interview with me, and I really enjoyed hearing about what you're doing, and the problems you're solving, and the way you're attacking it. I think it sounds like you've built something really interesting there, and I thin that the future in this space, we're going to see a lot more of this. I'm confident of that. But before I say goodbye to you, can you tell us where we can go to learn more about you and what you're up to?

Clint:
Yeah, just SaferMe. So safer.me. And look at our services there, that's probably the best place to start I suspect.

Daniel:
Excellent. Thanks so much for your time, really appreciate it.

Clint:
No worries.

Daniel:
And that's the end of another episode of the MapScaping podcast. My name is Daniel, and I really want to thank you for tuning in each week, it's greatly appreciated. If you want to reach out to me for whatever reason, you can find me on MapScaping on Facebook and Twitter, Map View on Instagram, or swing by MapScaping.com where we also have a full transcript of each episode of this podcast. Talk to you next week, bye.