Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Jesse Sprague: When the World Burns
25 Plays8 days ago
In this episode, Eric talks with Jesse Sprague, founder of EchoSpectra, about why wildfire risk isn’t primarily a firefighting problem, but a data, process, and decision-making problem.
Jesse’s work sits at the intersection of geospatial science, field investigation, and real-world accountability. What begins as a conversation about wildfire mapping quickly expands into a deeper examination of how organizations handle risk, why more data can sometimes increase liability instead of clarity, and how many industries quietly avoid seeing what they are technically capable of measuring.
They explore how wildfire investigators, insurers, utilities, and governments have historically relied on fragmented tools, handwritten notes, and disconnected systems and why that breaks down as fires become more frequent, more destructive, and more legally scrutinized. Jesse explains how EchoSpectra helps teams document fire behavior, fuels, and origin-and-cause evidence in ways that are defensible, collaborative, and scalable.
Along the way, the conversation touches on pipeline safety, change detection, insurance economics, smoke as an unaccounted public health cost, and a recurring theme: ignoring information doesn’t eliminate responsibility. It only delays it.
This is a grounded conversation about risk, systems, and what happens when reality outpaces the processes designed to manage it.
Topics Covered
- Why wildfire risk is a data and process problem, not just a response problem
- How investigators document fire origin, spread, and behavior in the field
- The difference between collecting data and being able to act on it
- Why organizations sometimes avoid visibility to reduce perceived liability
- How geospatial intelligence changes wildfire prevention and litigation
- The hidden public health cost of wildfire smoke
- Parallels between wildfire risk, pipeline safety, and regulated industries
- “Choosing your hard” as a decision-making framework for leaders
- What scalable risk management actually requires in the real world
Episode Links
- Connect with Jesse on LinkedIn: https://www.linkedin.com/in/spatialdata/
- Check out Jesse’s company: https://www.echospectra.com/
- Email Jesse: Jesse@echospectra.com
For more episodes: https://unfoldingthought.com
Questions or guest ideas: eric@inboundandagile.com
Recommended
Transcript
Jesse's Background and Current Life
00:00:01
Speaker
Jesse, thank you for joining me. Where does today's recording find you? Eric, ah well, at my desk. I'm feeling pretty ah pretty good. I don't know, relaxed or tired, or or maybe a bit of both, but had a good workout in this morning, and and I've got kids, and so kind of ah loving life and and doing all that. But otherwise, North Texas, five years into a a startup, and...
00:00:27
Speaker
um getting ready for a a weekend with with the family. That's that's where today is. Tell me a little bit about yourself, Jesse. I grew up in New Mexico.
00:00:39
Speaker
i my My father was a design build engineer for Intel. So he designed and built the fabs that manufactured chips, um you know, in the ninety s and And so grew up with computers in the house, ah which I think played a role in in kind of the the path that i ah wound up choosing or traveling.
00:01:02
Speaker
um And then my mom, she she's a pretty like outdoors type person. um Grew up in Colorado, spent a lot of time in Durango. And so I've kind of got a a an interesting ah combination of exposures there.
00:01:21
Speaker
rode horses. I kind of grew up a little bit rurally. so So on the outskirts of Albuquerque, we had horses, sheep, goats, chickens, things like that. And um ah it was ah an interesting combination because there'd be times when and I just wanted to, you know, take apart a computer or or explore the internet when it was like not great back then. You know, you're typing, just trying to figure out what you can find and click links until you wind up in some FTP site. And you're I don't where did this come from? Oh, that's an unsecured government site. All right, cool. What else? um Back in the day, but then also taking breaks from that and just being like, i you know, I don't want to be inside. and And that was kind of my mom's rule. Like the sun is up, you guys are outside. And so, um you know, a lot of time on a horse and which I and i haven't ridden horse since I graduated high school. But um that's kind of
Educational Journey and Career Transition
00:02:09
Speaker
my background. And then I went to college and thought I'd be a mechanical engineer and quickly changed. I realized that that was
00:02:19
Speaker
only partially interesting. and And I liked the kind of the physics side and, and, and I liked the scientific method more ah than kind of just the designing things. And so I, um, hi kind of went on that whole find myself journey and, and,
00:02:34
Speaker
thought I was going do a major in photography and chemistry, ah which was an interesting precursor. and And I've told you this before, Eric, was is ah i I really liked working with image data. And the the um chemistry side was just also very interesting to me. And so I wound up several years later getting a job at the United States Geological Survey. And I worked for a geochemist.
00:02:57
Speaker
And they, as you might know, may or may not know, the USGS is also responsible for maintaining all of the satellites, the public domain satellites that image of the earth. And so I kind of got my photography exposure and my chemistry exposure all at the same place while making maps, which I've always you know enjoyed. And I think partially because of of the outdoors that I was able to to experience as ah as a youth.
00:03:21
Speaker
And i and that that kind of precipitated where I think I wanted to go. And so I was i graduated with bachelor's degree in earth and planetary science from the University of New Mexico.
00:03:33
Speaker
worked for the USGS. Then I went to the private sector and did a lot of um groundwater remediation, which was fun, but repetitive after a point, just cleaning up oil spills and pipeline releases ah in in northern New Mexico.
00:03:48
Speaker
And so I went back to went back to grad school and um looked at a combination of computer science and geography. i so you know wanted to use computers to make maps. It's been a and of a lifelong passion.
Founding of ECHO Spectra
00:04:00
Speaker
And i kind of decided or or learned that I'd given myself a pretty decent computer science education just as a, as ah an enthusiast, kind of a hobbyist.
00:04:13
Speaker
Uh, and so, so transitioned to uh, actually ah an overlap between the business college and the geography college. And so, um we put in a proposal, my, my advisor and i and we went through a program with the national science foundation to look at the commercial is the commercialize ability of some research that he and I had done, um, the commercial potential.
00:04:41
Speaker
And uh, thought we would work in the oil and gas industry with some remote sensing and change detection technology. Uh, and, and so we interviewed a bunch of people and, and,
00:04:52
Speaker
ah That was in 2016, 2017. twenty sixteen twenty seventeen And about two years after that, I ended up starting, I wedged a lot of kind of consulting work to figure out where this was going to go.
00:05:06
Speaker
And then in 2020, I formed the company as it stands today. And i the rest is history. Now we we manage spatial data. We do a lot of, um we transitioned to kind of wildfire ah focus a couple years ago. And and that's what we've been doing.
00:05:22
Speaker
And the company as it stands today is ECHO Spectra, correct? Yes. Technically, that's not our registered name. We're the Axon Intelligence Company. who I kind of founded it on the idea that we would do geospatial intelligence and that that we were kind of analysts. um ECHO Spectra is our our brand name and our kind of our public facing image, if you will, and the name of our our software. So if you went to ECHO Spectra.com, you would see essentially everything that that we're doing. There's nothing under the the Axon um label. But that's probably nuance that's it's not totally important.
00:05:59
Speaker
ah So now Echo Spectre is our company. And and so we have ah an app for um an app for mapping really wildfire related phenomena, either pre or post fire.
00:06:11
Speaker
ah but But we've got some people who are using it just to document anything in, you know, it's offline compliant. You open your phone up in the fields, you drop a pin on the map, attach some photos and notes, and that's available to, you know, to you when you log in online, but also to anybody else that's on your team, they can see that information.
00:06:30
Speaker
um And so, so it's useful to a couple of different applications, but we really focus on on fuels documentation. So, ah hi are, you know, is this a ah grass or a shrub or or some combination of which, and there's some technical um categories there that impact fire behavior. And so we'll, we'll work with people to map those to try to prevent fires, um to help with what's called fuels treatments plans. be Like, all right, you need to thin this, or you need to mow over there, or you need to, you know, cut these trees down.
00:07:01
Speaker
but then also after the fact, so there's a thing called an origin and cause investigation. And as a As wildfires are increasingly costly, as we've seen just even over the past two years, there are four and six and $10 billion dollars plus um claims against companies for starting a fire.
00:07:21
Speaker
The investigations that are used to ah identify the cause are increasingly scrutinized. And so we work a lot with investigators to create a defensible,
00:07:35
Speaker
um you know our app allows them to collect a lot more data and in the same amount of time to create a much better picture and also kind of increase the well really just increase the defensibility of what they're conclusion is.
00:07:52
Speaker
I think what I've understood from some of our conversations previously and what you're saying here are is something like we have overhead photography, satellite-based or whatever, perhaps, and we're looking at an area of brush or forest or grassland or I don't know what that might be adjacent to a community or it could be adjacent to some business I don't know what and there's some amount of perhaps in advance of you know some sort of wildfire
00:08:32
Speaker
documenting and understanding what the wildfire risk is so that perhaps we need to do some management of that land, clean up brush shirt something.
00:08:43
Speaker
But then also, if I heard you correctly, there's also after the fact understanding where did the fire start and there may be more to it. So I i suppose bringing you back to a question, i think the question really is,
00:08:59
Speaker
What is it that people are doing with Echo Spectra or with your work generally and and why are they using it? guess the initial kind of important thing to to ah to clarify on on my part is is we have the software that allows users to document things from their customers.
00:09:18
Speaker
their phone or their iPad or or tablet ah in in the field or or wherever they might have that device. So there's an app. um They can look at the data online. and And that's one thing that we will dig into, kind of the fieldwork component.
00:09:31
Speaker
The other thing that we do is we call it professional services. And so we we analyze and integrate tons of different data sources to reconstruct ah usually post event.
00:09:44
Speaker
um But sometimes we do proactive for like risk assessment. We will take we will take that fuels data, whether we collected it or they collected it, or if it's a different source of fuels information, we will look at weather forecasts, we'll look at construction locations. For example, somebody wants to build a um an apartment complex or ah of a large HOA that is, you know, ah a single entity, you know, they say from the studs out is owned by one company, they need to get insurance. And so we'll run simulations against that location to say, okay, well, here's where your fire risk is, here's what you can do to mitigate that risk. um That's a professional services thing. So that's not something that you would do in the software, at least not yet, you know, maybe someday.
00:10:28
Speaker
ah But then also after the fact, we will take I mean, information from places that i i didn't know was available even a couple of years ago. So we'll look at, um and you know of course, people are familiar with the fact that your phone is tracking you. And so um we've been involved in in legal cases where that becomes discoverable. And so our our clients will say, you know hey, I've got this thing from AT&T.
00:10:57
Speaker
there's probably some information in here, but to us, that's a gold mine because it's just a bunch of timestamped locations. And so we will create an animation, say here's the the replay of what that data set tells you. And we do that for just a ton of different types of data sets. And then we layer that on top of, you know, satellite data and thermal anomaly detections from satellites. then we have some kind of cool things that increase the um the resolution and the confidence intervals or help to refine the confidence intervals around where that thermal anomaly is being detected.
00:11:28
Speaker
And then when you look at that in ah in an animation, like in a time series, all of a sudden stories start to come out and you you you you're armed with this ability to ask questions you didn't even realize were you know questions that you should be asking. And so so um we we do quite a bit of that on the professional services side.
00:11:47
Speaker
and And so data integration, data science, a lot of data visualization, but almost always around fire, either fires that have happened or fires that you want to
Functionality and Impact of ECHO Spectra's App
00:11:55
Speaker
prevent. and um And so you could think about, you know, one of the things that a great kind of anecdote is,
00:12:06
Speaker
You have this big field of, you know, knee-high grass that they say cures out in the fall. It's dry. It's a very flashy fuel on a low humidity, like ah a low relative humidity day. ah Somebody, a truck could be dragging chains or somebody could throw a cigarette out the window and it will light that grass on fire.
00:12:24
Speaker
Catalytic converters start fires all the time on on you know really dry days also. And so we um there are people who are like, well, just burn all that grass down. That's a fuels treatment. If you do it controlled and don't let it get out of hand, then that fuel is no longer there and it can't blow embers or or fire can't progress through that, gain momentum, and then start a stand of trees on fire, which becomes much harder to then to them put out.
00:12:48
Speaker
The problem with that is that when you when you do ah ah a fuels treatment plan on grass, it just grows back next year, right? And then like it might even be worse because you've basically gotten rid of anything that takes longer than a year to grow. And so things that would be more fire resistant um aren aren't there anymore. And so you're just stuck with a problem.
00:13:09
Speaker
kind of a down or a worsening problem where you have flashy fuels, you've got to get rid of them and then it comes backwards and, and, your mitigation plan, uh, was not resilient, so to speak. Um, and so we can help people say, okay, well maybe what you should do instead is, you know, some, you know, mow this thin strip or just, you know, turn, you know, turn this, this thin strip over. So you have a fire break, but then let's do some thinning in the trees so that they're less likely to catch, remove ladder fuels and things like that. Um,
00:13:38
Speaker
So, so we, that's kind of the proactive side. Um, and this, again, that's all on professional services. And then to the other, uh, part of your question, the the, the, the mobile app, which is used to collect that data in the field kind of serves a couple of, a couple of, a couple of purposes. There's, um,
00:14:02
Speaker
fuels teams who will take that app into the field field and say, Hey, here's grass. Here's, ah you know, ah a category of fuels that is characterized by a lot of ground fuel. That's flashy. Like maybe there's grass and and shrub, but then also a lot of we call ladder fuels, stuff that can, the fire can climb and get into a, into the crown of the trees. And they call, they call this like a, a transition or ah a ground to crown transition.
00:14:29
Speaker
And when that happens, your fire behavior becomes, um, much different, much more extreme. And so it's a lot harder. It's a lot more risky. So a crown fire, and, and, and this is actually counterintuitive to me, but when a, a dead tree is burning, even if it's all the way up at the top of the tree, the whole tree, they say fully involved. I, I don't really have a background in fire. So all the time I spend with these investigators and firefighters, I pick up the, the words they use. I'm, I'm a data scientist and a a geologist, really. But um the i that though the technical term being um fully involved, a dead tree, which sounds like a lot of fuel and it's like really hazardous, is actually not as bad as a live tree.
00:15:13
Speaker
Because when you get a live, for example, a pine tree fully involved, there is a lot of moisture in it and a lot of oil. And that oil literally just explodes. i mean, you see pictures of these live trees that have been heated and are are on fire that, you know, they're basically just exploding. They're throwing these embers everywhere. And that ember can start a new fire downwind, they'll ember cast. And and so, um or spotting downwind. And so you'll get you get a new fire two miles away or three miles away. And so you can imagine the firefighters are trying to respond to ah ah a front, trying to you know curb or do what they can. And all of a sudden they've got a new fire behind them that they don't even know about yet. And it's moving that way as fast as the one that they're trying to stop is moving towards them. And that's really dangerous. And so especially if there's a change in wind direction because they don't know
00:16:04
Speaker
you know, that there's a fire behind them. And so, so, um I guess all of that's to say is that these fuel crews will go out and document be like, Hey, there's a lot of ladder fuels here and we don't want this to become a crown fire. So let's, um, we want to propose to the County or the city or this neighborhood that they do this kind of work. And there's a lot of, of grant money out there for them to do that.
00:16:26
Speaker
And so they'll collect that data and then we will then take it and use that as our labels in a satellite image, uh, processing, i guess, process. so So we will use the ground truth that they collected in the field, train a model, and then categorize every pixel in the image based on the couple that we know, you know, we're like, hey, this pixel, they have photos and documentation on the ground. We know this is this category of fuel.
00:16:53
Speaker
find every other pixel that satisfies, you know, approximately the same condition. And that's probably where those, that fuel type exists across this area. And so, so that's one of the ways that the app is used and then kind of in turn becomes a way that we can leverage that data for, for our customers.
00:17:09
Speaker
Excuse me. um The other, and, and, and the majority of the app users are fire investigators, fire investigators, and they can be law enforcement or kind of a private, uh, consultant type company.
00:17:28
Speaker
And generally the, the consultants are hired by insurance companies or law firms or law firms that are representing insurance companies or power companies. Uh, of course, plaintiffs also hire their own investigators in these type of, of cases.
00:17:48
Speaker
Um, And so there's there's a very legal component to that. ah there's it While still a legal component to the law enforcement, it's essentially evidence collection. So if a fire crosses a property boundary or burns down a building or several buildings, or if anybody's died or even injured, that becomes something that a a fire investigator who works for the county or city or state, ah who is technically a law enforcement officer, goes out and investigates.
00:18:15
Speaker
And as you can imagine, there are a lot more law enforcement officers than there are, uh, wildfire events. And so they, they generally don't have, or didn't before have a really good system to document in a, in a consistent way. And so we, through our work with all of these people, we're like, Hey, we we can help here. And so we, we, um,
00:18:39
Speaker
created this app and it follows the National Wildfire Coordination Group, NWCG's guidelines for investigating fire. And that's called specifically an origin and cause investigation, which is funny because even just a couple of years ago, they called it cause and origin. Everybody got would get kind of worked up. They're like, you know, you can't have a cause before the origin.
00:19:00
Speaker
That's just bad investigation, right? It's bad science. And, and, uh, and there's, you know, they'll say like, oh this is a cause in search of an origin. And so, you know, a plaintiff or, or somebody who was upset might go out there and say, well, the fire started somewhere over there.
00:19:17
Speaker
Who in that area has the deepest pockets? It's their fault because we need to recoup the cost, which is not the right way to go, but it happens. And so even in so, so the,
00:19:29
Speaker
it's been the wild west and it still is to a point, but we're finally getting some science around the process and, and some structure around the process. And we've had a lot of fun working with these investigators to, to kind of shore up the defensibility of their cases because, you know, 10 years ago, a really big case with $300,000, right? I mean, yeah, you burnt down half of my pecan trees.
00:19:53
Speaker
I'm going to get what I can out of it. And and that's it. But that's not the case anymore. that Like they're, like I said, $10 billion dollar claims against power companies or or other other entities for causing a fire. And then anybody who might even possibly be involved um gets sued it went in you know to try to, ah they call it cost recovery. It's it's um it's become a much bigger number, which means that the investigations are under a lot higher scrutiny.
00:20:22
Speaker
um And then as we start to to, I guess, work with them, and I should probably round out the you know that one specific example there, Eric, is is i as they, as they do that investigation and they document what they're seeing, which are called fire progression fire progression indicators, they'll be in the field, they'll see a rock, and it'll be like black on one side and not on the other. and they'll call that sooting. And they'll be like, because of the way that the sooting is oriented, I know that the fire progressed this direction at this point.
00:20:58
Speaker
So before they'd go out and they document that they'd plant a flag, they'd take a GPS coordinate, they'd take notes, they'd take photos, they'd basically be recording all of this. You know, they've got their handheld GPS, they've got a, you know, a DSLR camera, they're right in the rain notebook, ah you know, oh, it's a mess, right? And then they got to get they've got to get back to their office. They've got to figure out which photos, go to which entry, plug it all into a spreadsheet, turn that into a KML, look at it in Google Earth, try to remember which one was, and are different types of of behavior. For example, you have advancing fire. You could think about it. If you just drop, ah you know, a match in the middle of a bunch of grass,
00:21:38
Speaker
even if the wind is blowing in one direction, the fire still grows laterally to the wind. We call that, you know, aye lateral indicator. And it'll also creep upwind just because the fire wants to grow and consume fuels. And we call that a backing fire because it's it's it's it's crawling up into the wind, whereas the primary direction of of growth is downwind, and that's the advancing. And they all have different indication. They leave different marks on the ground and and on vegetation and and things like that, different characteristics.
00:22:09
Speaker
And so then they get there and they get, you know, they're back to their office and they're trying to figure out what they collected and it was a mess. And so you'd have five investigators and they would spend all day and they'd collect, you know, 50 points or or, you know, maybe a hundred points. And so they're trying to, you know, answer who's responsible over days of investigation, but one rain and half of those indicators are gone. And every day,
00:22:36
Speaker
that stuff washes away. It's just ash, right? And they're very ephemeral. And so we went out with several of these investigators on multiple occasions, and we we started using ah this app that we had built for collecting data in the field for this specific use case.
00:22:55
Speaker
And right out of the gate, especially if if there's good cell service or even decent cell service there, because everybody who's on the same team can see the indicators that everybody else is mapping, right? I can be walking up one canyon and be documenting a certain type of indicator, and you could be walking up the the adjacent canyon and documenting things over there, and I can see your information on my tablet in near real time or on my phone.
00:23:20
Speaker
almost immediately. And that allows the team ah collectively to make decisions about where they need to so focus resources. But they're also collecting like an order of magnitude more data over the same amount of time. And so they're able to create a much more comprehensive picture that would still go out and map for a week because, you know, they're pretty, pretty important that they get it right. But they're, they're collecting 10 times as much information, and which allows them to create a much more defensible picture of how the fire behaved, um which, which we then can help them do additional things. But a lot of times they'll just export that and say, Hey, here's our report. Here's our photos.
00:23:58
Speaker
You know, that's the map that they need to turn in. Who, ultimately benefits from either the professional services side or the app side or both. and And not just who benefits from, but also who's the customer.
00:24:13
Speaker
You know, is it me as a homeowner, insurance companies? Is it a big property owner, state, the county? If you are trying to get insurance, especially if you're in California, Colorado, or Utah, um and and your insurance says you're in a high risk area and you're like, well, is it or is there anything I can do to help um mitigate that risk?
00:24:36
Speaker
and And yes, you would hire us and we would run an assessment and we would say, OK, here's your primary risks. We have the credentials to make that a very believable assessment. Like we we have confidence that we're we're right. um And then here's the things you can do that are known, um you know, kind of the the highest impact, or the lowest hanging fruit, so to speak, of what you can do to protect your home. And and sometimes it's, you know, they talk about burn fire beds or burn beds, you know, don't have any vegetation touching your house.
00:25:04
Speaker
Make sure the vents under the eaves are ember proof, you know, that type of stuff, which, um you know, we've got people on the team who are are doing that for their entire career that they can help with. Um, that's not the majority of our work. Um, the, the people who benefit primarily, um, from the app itself are going to be the fire, like the fire departments, like, right. The, the fuels teams, the investigators, the fire marshals, the state and local governments, they get the most value out of the app, just all the cart because they are,
00:25:41
Speaker
you can think about it. Um, I've got a colleague who says, you know, I've never seen a fire start at 50 acres, but they put a lot of fires out before they get to 50 acres.
00:25:55
Speaker
And, especially on the West Coast and, and you know, in those forested areas in the Northwest, there there is a lot of missed information. And so one of the things that they're looking at is, can we catch arsonists a season or two earlier before it becomes a a major, you know, ah a catastrophe or, you know, it just just a disaster with loss of life or or very expensive, um you know, property loss.
00:26:23
Speaker
Because firefighters are generally pretty good at their job, they will show up, put that fire out, and then they'll go move on. They've got something else to do. um they've got you know they're They're busy.
00:26:36
Speaker
They're not going to go through the process of planting a flag, looking around the fire. like They just put it out. They're not going to document it for another day or two. But if... And and you might...
00:26:50
Speaker
Well, let's take Cal Fire, for example, you know, about 100 full-time investigators who are law enforcement officers. But then there's 1,000 or 2,000. I mean, there's like 10,000 people Fire. And there's another 2,000 like bosses captains
00:27:01
Speaker
two thousand i like like engine bosses or captains who are responding to fires all the time. And if they can document in a fraction of the time that it would otherwise take just what happened, it's like, hey, i mean, just drop 10 pins, take some photos and fill out a couple of of, you know, bits of information so that the organization as a whole can see kind into the fire problem and in a level that they couldn't previously. And that will start to identify trends and patterns, which could be indicative of maybe just something like, hey, we should really just mow the grass next to this road because there are always fires there. Like every season we have multiple of them. Perhaps we should allocate some of our vegetation management budget a bit more strategically and and and just cut that problem off upstream.
00:27:52
Speaker
Or they might say these fires are really, there's something kind of funny about them. They might indicate arson. And if you have a serial arsonist, um they're you know they're going to escalate, right? And so if you can catch them before it becomes a major threat.
00:28:09
Speaker
you know, a major disaster, then that's a win for everybody. So even you as a homeowner um have have won there. So stakeholders are really all of these communities. And and the the people who use the app are are generally investigators or fire professionals. um we i We are contracted by insurance companies,
00:28:33
Speaker
landowners, but generally not like you and I landowners, um, somebody who's got 10,000 acres or, or a, you know, ah really nice piece of property in Vail, and they want to build a very nice house and they're only there for a couple of months out of the year.
00:28:51
Speaker
What do they need to do to show to the insurance that they don't need to be paying a million dollars a year for their premium? Um, So that type of homeowner, um, we, we've, we get contracted by occasionally, but, but a lot of it is, um, law firms generally somebody representing, uh, think by by dollar amount, we do more defense work. So our clients tend to be like power line companies, insurance companies. Uh, but that's not, that's not, you know, exclusive. We've, we've been contracted by plaintiff, um, parties,
00:29:25
Speaker
Also actually sometimes insurance because the insurance is like, no, we're suing them so that we don't have to pay out. And so, um, uh, yeah, insurance and infrastructure are the kind of the, the, uh, was professional services type customers?
Challenges and Innovations in Data Management
00:29:40
Speaker
You said that when you got your business going or something like that, that, I think it was based on some research you had done. And so I don't know exactly how to phrase this, but I guess I'm wondering...
00:30:04
Speaker
what is it what was your research and then if it relates i suppose what were people doing before you provided these services and created this app or maybe what is the alternative if it relates to research the fun story about how we got here and the research that I was doing with, uh, with my professor and, uh, currently he's, he's now a technical advisor and has become a friend, um, since then what was change detection. So, uh, because I did so much pipeline remediation work, uh, on, on shallow aquifers, I was very familiar with the impact that the, the, you know, pipeline companies were having on, on drinking water and, uh,
00:30:50
Speaker
and kind of, i guess, shallow water resources in general. And every time I'd show up on site, you could see exactly where the leak was. But you're out in the middle of, you know, for me, it was the San Juan Basin in northwest New Mexico, where there's just not a lot of people out there. And so some places, you know, that leak was basically undetectable by the SCADA systems, the pressure monitoring and everything else on the pipeline infrastructure.
00:31:14
Speaker
But then some guy drove out there and was like, whoa, there's a bunch of dead vegetation or or holy cow, I can smell something or, you know, and there's some that are worse than others. Like you know, and ah an Arroyo is like, I showed up on site one time after a rain where the side, you know, a 15 foot wall of basically sand had fallen over and it just ripped a pipe right, you know, in half. It was like an eight inch pipe that just was sticking out in the middle of the Arroyo. And, um, they knew that one of the problem when they were able to send somebody out and shut the, um, shut the pipeline off. But you know, the, the, the tragedy there is that
00:31:48
Speaker
you know, and I don't think it's a concern, I'm not going to tell you who it was, but thousands of gallons of hydrocarbons were pumped into just, you know, spilled out of this pipe into a wash, which fed directly into the the local river.
00:32:06
Speaker
you know, the wash, there was water in it while it was, um you know, during during the flood or the kind of the the storm. ah Flash floods are really common in that area. But 20 minutes after the storm was over, the water's gone above ground.
00:32:20
Speaker
Below ground, there's still a lot of water moving through the sand in the called shallow alluvium. And so so i the contamination hit that kind of shallow groundwater and was gone there. We had no chance, but I mean, we showed up two or three days later and we couldn't even, we were sampling hundreds of yards at a time trying to catch the contamination. of course it's diluting as it goes. And eventually it just, you know, reaches an equilibrium of of a lot of dilution, but that's, that's, um, your highly diluted, uh, state, but it's, um, still pollution, right? It's still in the groundwater and that's water that's going to wind up
00:32:58
Speaker
interacting with the entire, you know, ecological system along the with the the riparian ecosystem for the, you know, the next 50 miles or more at ah at a level that matters. and And so anyways, as as an aside, um the the every time I'd show up, I would say, i can see what was going on here.
00:33:17
Speaker
I could have seen this with aerial imagery or a drone or or sometimes even a satellite. and And so I had this idea to monitor these corridors, which they're required to do. So they're they're required to inspect them every so often, depending on the, you the risk level. You know, if it if it's a 30-inch high-pressure pipeline that's running through it a neighborhood, they're probably looking at that weekly. um But these, you know, 10-inch lines that are running through BLM land in the middle of nowhere, um they probably only see that quarterly or less, um maybe maybe twice a year.
00:33:51
Speaker
And there's rules about this. The Department of Transportation, FEMSA, have you know, there's the Pipes Act of 2020. It's fairly well codified, but it's a guy in a plane looking out his window for something that catches his eye. And and sometimes, or you know, a guy in an ATV driving these corridors.
00:34:10
Speaker
and And sometimes they see things, sometimes they don't, but it's it's ah infrequent. And so we developed this change detection technique that was very efficient and very fast and runs on a GPU at a time when people didn't have any idea who NVIDIA was, except if you weren't a gaming nerd, you didn't know what we were doing in college, like from writing code to do, but... you know, that's changed.
00:34:29
Speaker
Um, so, so we're writing code for, for these NVIDIA GPUs to do change detection in a, in an efficient and, uh, you know, it's kind of scalable, very fast way. And we went through that and told about the NSF program. We went through the national science foundation funded ah our kind of customer discovery. And so we traveled around and we talked to all of these different pipeline companies and they, um,
00:34:53
Speaker
they loved what we, what we could show them. I mean, we were detecting tire tracks on the right of way between aerial patrol visits and, And so they asked us to go, ah they asked us to go to, to, to do a demo. And so we did. And as part of that demo, we mapped, you know, a 200 mile stretch of pipeline in North Texas for ah once, i don't know, like April. And then again, in maybe June. um And we looked for all the change and there was a lot, quite a bit.
00:35:20
Speaker
And they loved that too. But there was enough that somebody asked, you know, how much are we getting currently? And this was the big learning for, for not for just for us, but for for our client also, um the people who are doing the aerial patrol were turning over 2,500 observations a year, maybe more.
00:35:37
Speaker
And the customer only, they could only pull up three or 400 a year. And so if they, if anything had gone wrong on the other 2,300 2,2100 of them,
00:35:51
Speaker
an an An auditor would come in and say, you know, give me all the information that your aerial patrol produced. There's your 2,500 locations. Well, if any one of those locations are near where the accident happened and they don't have a documented response, they're not just liable, they're negligent.
00:36:09
Speaker
And that's like a really big problem ah for them, right? Like people go to, your vice presidents go to jail for that. um And, you know, people die in these cases and and that's happened. um You know, pipeline explosion is pretty serious.
00:36:20
Speaker
And so then here we show up and we're like, I can take your exposure liability and I can 10 exit because I'm about to dump another 5,000 or 6,000 observations that you need to know about on your desk. And they didn't even have a system to manage what they were already getting.
00:36:36
Speaker
And so that was our realization. My my colleague and I were like, OK, we're a little bit ahead of the industry. We have a really cool technology, but there's a process problem. And and it's not a, you know, it's not ah a technology or capability problem because the guy in a Cessna is doing good enough right now.
00:36:54
Speaker
But there's a there's a process problem because there's even that is creating liability for him. So, so the analogy that that I've used and several times is is that we had a lot of mail, but they didn't even have a mailbox. And, and so the, you know, it was just piling up on their door and they were missing bills and everything else like that. So there's a problem.
00:37:13
Speaker
Does some of this look like, so a guy in a Cessna, he looks out the window basically. And he says, oh, I noticed this thing. And he writes it down or maybe he takes a picture and it's perhaps not going into a computer. It's not going into a system where someone can be notified.
00:37:33
Speaker
This might be higher priority than the next thing. And then if it's something like that, then a kind of a first level response might be, oh, well, we need to develop a system where we can categorize or we can prioritize things. And so the is that what some of this challenge looks like?
00:37:55
Speaker
And then if so, you know, were you presenting such a dramatic growth in the data that even if they had a way to prioritize that there was just too much, they were not staffed or did not your clients or potential clients would not have the capacity to deal with all of these, these data points. Yeah.
00:38:20
Speaker
Absolutely. So we call that risk ranking. Right. and and And the way they handled that currently is that pilot, you know, if he saw a release, he'd just pick up his cell phone and be like, hey, I'm on such and such line over here, at such and such. And I see something you should send somebody out like right now.
00:38:37
Speaker
go take care of that. And so for fires, obviously, like if there's somebody, you know, burning or if somebody's doing a ah like, like if he sees yellow iron, like like ah somebody's plowing or digging or something like that in the right of way, he'll just call, right? Like that's a high risk. Somebody's about to hit the line and blow stuff up or or something like that.
00:38:55
Speaker
Um, but all the other things like somebody built a barn or somebody built a play set or somebody put a runway, like the private runway ah across the, you know, pools are very common, but obviously very dangerous. Even planting trees on the back of, you know, your property fence, your property line.
00:39:10
Speaker
If it's in a right way, you've got to dig a hole to plant the tree. Like all of those are the type of things that are like, it's already been done. They didn't blow up while they were putting the pool in. It's unlikely to blow up now.
00:39:21
Speaker
You should get somebody out there and, you know, much to the chagrin of the client, of the customers, the the the guy who put his pool in, tell him he's got to move his pool, which, um you know, he he would have known if he tried to get a permit or anything else. But, um yeah, so there is a risk ranking system kind of in place with the status quo, but it's not scalable.
00:39:41
Speaker
And there's a lot of things that that pilot's not going to see, especially if he's just trying to remember what did he report six months ago when he flew this same line. And yeah technically his job is to report it anyways.
00:39:53
Speaker
You know, he'll report the same pool every time, but um you can see how how I guess that the, the data coming in is very similar to the problem we had with wildfire. And that's actually how we built the foundation of the product that we're using, like this whole concept of handwritten notes, you know, a GPS point and a camera out of the, you know, out of the window of a plane. It's very similar to what was happening in these wildfire investigations.
00:40:21
Speaker
And so fortunately we had had some experience with that, that kind of, problem already. But then on the other side of the problem was that the client the the pipeline companies didn't have a way to handle even the data that the aerial patrol was sending them. Because they'd get a an Excel sheet and an FTP link and then they would save all of this data but it would just be on some guy's computer. And then that guy would get a promotion and all of that historical information was gone. It's not documented in the database anywhere.
00:40:50
Speaker
and if And if it was an emergency type thing like on a shorter patrol or a more routine patrol, The guy in the CES or the, you know, the aerial patrol company would send an email like, hey, we flew this 50 miles today. Here's three things and here's the photos attached.
00:41:03
Speaker
Well, the pipeline companies have their settings on their exchange, their Microsoft Exchange server. So those photos are gone in six months. And so they can't pull it back up later on to to even figure out what was what was happening. So they they had a process problem internally just trying to manage the data that was coming in.
00:41:21
Speaker
And so not only could they not risk rank it, but they couldn't look for trends or patterns. Like, for example, this area is subject to a lot of development. They're building neighborhoods like crazy. Well, that's going to change their risk profile for that pipeline that's in the area. They should probably change their patrol frequency. They should probably pay extra special attention to this and the other thing. Like, we found things on multiple occasions that had been reported several times to the clients.
00:41:48
Speaker
for the over the course of months. And then all of a sudden it's like they reach out and they'll say something like, hey, ah there's an apartment complex on top of our line.
00:42:00
Speaker
Like, it's like, well, yeah, you know, we told you about that for months, right? I mean, here's that here's the eight reports that you got that construction was happening from when they first broke ground to when, you know, all of this was happening and and they just didn't have a way to respond to that because, and again,
00:42:16
Speaker
They weren't seeing the data in context. They'd get a report of something in that area. And then three months later, they'd get another report and, you know, or a week later, they get another one. And and there are these points that are coming in over time. And they're all on their own fairly low.
00:42:31
Speaker
ah Well, until you have an apartment complex, they're fairly low, I guess, risk. But when you look at the map and it's like, hey, we have six of them in the same spot. now they they're potentially interactive, right? And so there's like a um this this spatial component that says, hey, it's not just like six spread out over that 50 miles. They're all in the same location. You should you should probably respond to that.
00:42:54
Speaker
In the the experience that you went through, i think it was with and NSF, when you you know went out to some potential customers, the response essentially was We can't do this because we'll have too much exposure?
00:43:12
Speaker
was that Is that it? We were going to give them more data than they could respond to. that Now, they the catch there is that as soon as they knew that they had a documentation problem or a process deficit, they were responsible for doing something about it. So they had to come up with something.
00:43:29
Speaker
And that's where we shifted from being this aerial based change detection kind of provider. Like, hey, we will collect and process this data. We will send you pins on a map that you need to know about. And that's what you're going to subscribe to is essentially ah a a digitized um aerial patrol, so to speak.
00:43:47
Speaker
um we were And we were going to provide it with those aerial patrol partners and collaborate because there was a lot of value to doing that um in general. But they they realized in exactly what you're saying, they're like well, our liability is going to go through the roof if you report all of this stuff. But they knew that there was more going on that they needed to know about. And so we took, and I said 10 steps back, and we built that mailbox. We were like, okay, here's a system that allows you to document points that represent risk or change or anything else.
00:44:19
Speaker
And you can risk rank them. You can assign them to people. You can collaborate on them. i mean, in it in its simplest, simplest way, ah iteration, it was a lot like a Trello board for observations, except there was also a map view.
00:44:35
Speaker
This sounds somewhat similar to how pharmaceutical companies and even I believe healthcare providers to some extent will very often when they do advertising, you know, online advertising or whatever, they'll avoid channels where people can comment.
00:44:56
Speaker
So like if if a pharmaceutical company is advertising in Facebook, Instagram, whatever it is, they will turn off comments or they'll just avoid the channel altogether because if someone reports an adverse effect or something of that nature, they're, I believe, legally required to respond.
00:45:21
Speaker
Even if it's something that is completely off the wall and, you know, one of their researchers, doctors, whatever, might say, there is no way that our medication caused that, but they're required to track it down. And not that it's the same, but it sounds actually quite similar. Like, I'm just going to turn my gaze away from that data point. That way I don't have to deal with it.
00:45:48
Speaker
Yeah, no, it's, um I guess the saying there that, you know, the ostrich still gets eaten. He just doesn't see the line coming, right? Like, you bury your head in the sand all you want. It doesn't save your life, you know. You can pretend it doesn't happen. You're still going to die. Yeah.
00:46:04
Speaker
but maybe it's just less stressful. ah you know and But on the other hand, i can see the justification for, hey, we have a system in place that is working basically good enough, right? There's a lot of the medicals, FDA approvals and everything else and you know in the pipeline and power line, the required patrols that they they have to do.
00:46:26
Speaker
they're They're exceeding this currently set standards. And if we dumped another 10 times as much data on their desk, their existing processes would have broke. And so we we we would have made, we probably would have decreased their safety performance from the status quo even. And it's not just we would have increased their liability, they would have struggled to deal with it at the expense of what was working good enough. We'd have just broken the the current systems. so So even though the the medication is only causing that huge role of of side effects, they're doing a lot to mitigate and and already, you you know, I guess, prove that the medication is safe.
00:47:07
Speaker
And if we distract their team or if we make it way more expensive to do that job by forcing them to respond to, you know, everything that gets commented on their Facebook post, now their current processes, you know, they're they're their quality control is going to go down because they're, you know, a finite resource in that regard also. Or the cost of the medication is going to go way up.
00:47:27
Speaker
Is it fair for me to think that
00:47:33
Speaker
if a wildfire investigator, insurance company, HOA, whatever it is,
Future Directions and Personal Reflections
00:47:41
Speaker
is able to get data in advance or after the fact that we can end up in a world five, ten, whatever years from now where we have lower...
00:47:59
Speaker
you know wildfire rates, incidents, you know, numbers or whatever, and when fires occur, that the losses are less because we were better prepared in terms of, you know, mitigating whatever the risks were, maybe even we respond faster.
00:48:21
Speaker
Is that is that a fair assumption for me to make? Yes. And we're not the only people focused on that. That is, um, any of the conundrum here is that wildfire is very healthy for the ecosystem. It's a critical part of a healthy ecosystem.
00:48:34
Speaker
In fact, um, it's what we call the wildland urban interface, the WUI, where people start building houses at the, you know, the, the,
00:48:44
Speaker
into this ecosystem where fire is a natural and healthy component that you start to have have a lot of risk to the built environment, to you know all of the money you put into your house, to you know your family that sleeps there.
00:48:56
Speaker
um And we do want to mitigate that that risk. we We want to make it safer. And that's kind of our um thesis is that because we do so much post-fire documentation, we're uniquely positioned to provide very credible, defensible, and accurate risk assessments.
00:49:17
Speaker
Is it reasonable for me to assume that because it is so difficult for a pipeline company or for a large land owner or for Cal Fire, I don't know who, you know, for someone, an interested party, because it is so difficult for them to keep up on all of these potential risks that they the total dollar amount, you know, the total damage that has occurred and the, you know, settlements that are being awarded in court, that those numbers over time have grown perhaps in part
00:50:01
Speaker
to make up for the fact that, as you said, with a pipeline or whatever else, they're unless there's an apartment complex nearby, the risk that something occurs is relatively low.
00:50:15
Speaker
But because there may be a hundred minor leaks somewhere, or there may be a problem that occurs and no one is affected, but then when there is a problem and a lot of people are affected,
00:50:28
Speaker
I feel like it's a bit of a reach, but... Do you think it's reasonable to assume that the a settlement amount, for example, in that one instance might be kind of making up for all of the other instances in which there was no settlement to be had because there was no documentation? there was Nobody even saw the tree fall in the woods.
00:50:51
Speaker
It's an interesting idea. I think as an, ah what are we, they're not ecosystem, it is an ecosystem, ah an economy. I think the the economy of you know, fire impact and litigation, you know, like for example, a fire burns and they put out the flames, but we'll say, you know, five years later, that file, that fire is still burning. But at this point, it's just burning people's time and money, right? Like they're just trying to like settle it out. um I do think there is ah an equilibrium there um that you would naturally get, which is, know, general economic forces.
00:51:23
Speaker
And I think that's probably partially like, the perception, right? It's at the end of the day, especially if it's a legal case, it's being negotiated by people and potentially a jury and the, the, they say the court of public opinion, right? Like people know that there's, there's an impact, but you know, what's interesting, Eric, and then I think that you're,
00:51:45
Speaker
It's very funny that you bring this up because ah you know it ties into your medical thought, but but the the medical community idea, but um nobody is quantifying the health impact of smoke.
00:51:59
Speaker
And for example, ah fires in Alaska are left to burn. They basically, you know, the Forest Service or somebody will just send this one guy out there and I can't remember what he's called, but it sounds just like the greatest job. He's basically a fire scientist and then just post him up out there and they're like yeah, just watch it, radio in, keep us posted, you know, don't die. ah um And, and ah you know, they don't try to stop it because there's no people out there. There's not, you know, l LA right next to that fire. and And again, it's a healthy part of the ecosystem.
00:52:29
Speaker
ah But it is dumping a ton of air pollution, even though it's natural. I mean, we so there's a thing called norm ah in the, yeah you know, the Earth planet size, is normally occurring radioactive material, naturally occurring radioactive material.
00:52:44
Speaker
Just because it's naturally occurring doesn't mean it won't kill you. um And so it's an important part of the ecosystem and the environment, ah but it's still a pollutant. and And that smoke is bad. And it is just being funneled right down the jet stream into the population.
00:53:00
Speaker
However, even worse than that is the smoke that comes off of burning the eastern half of L.A., right? Like drywall and all the chemicals in, you know, your grandpa's garage and everything else, all the Ace Hardwares and everything else that just burnt to the ground.
00:53:16
Speaker
all of that contamination was funneled right into the neighboring communities and states. And nobody's quantifying that impact. So while maybe there's a ah you know this court of public opinion, like fires are causing a lot of damage. You happen to be on trial for this one fire. We're going to get as much out of you as we possibly can.
00:53:36
Speaker
i think that does exist. i think you're right to consider that. um However, i don't think anybody is really cognizant or or quantifying the total health impact um of of of these fires in a way that would change the, I mean, you know, who bears the burden of those costs, whether it's the country in just general health and productivity of its of its citizens, if it's, um you know, health insurance companies, if it's, i you know, it could be of any number of of stakeholders.
00:54:12
Speaker
But generally, as a you know, a community, whether that's a country or, or smaller, you want your members to be healthy and productive and, and smart and educated and, and you know, everything that makes your community more successful and, you know, dumping a bunch of pollution on them is, is counterproductive to that, that agenda. And, and and I don't think people really,
00:54:33
Speaker
um thinking they've grokked that just yet. ah well We'll see where that goes. Are there places that I should go to connect with you or learn more or things that I should go and check out? And then second question or second area of questioning, whether we've talked about it already or we didn't touch on it, are there words of wisdom or things that we should touch on or you would like me to be thinking about after this?
00:55:05
Speaker
Far be it from me to tell you anything that you should think about, but I, I, there's been something on my mind that I'll share, um, as a, an anecdote or, or, you know, closing remarks. Uh, you can get ahold of me. I'm on LinkedIn. That's probably the easiest way. as much as I, I don't like it. That's my, my compromise. I don't have any other social media. So, um, ah you can find me Jesse, Jesse Sprague on LinkedIn. I think my you know, slug or or name is just spatial data. So profile slash spatial data. Um, yeah, reach out anytime.
00:55:38
Speaker
Jesse at echo spectra.com. Those are probably the the easiest ways to get ahold of me. Um,
00:55:46
Speaker
something to think on food for thought. I had a friend, uh, well, still a friend and and a good friend. Um, and
00:55:58
Speaker
I've been thinking about something he said a lot recently. And so I'll share it because I think it's, uh, worth thinking about. He said, we're, we're training partners. Uh, he's an iron man multiple times over and I'm training for my first iron man. I've been a cyclist for pretty much my whole life. And so we spend a lot of time on the, on the road together, uh, just working hard. And, and that kind of stimulated some conversation around, um well, hard work and and things that are hard and how um you know adversity is is you know a key ingredient, if not you know the the primary driver for so much growth, whether it's personal or or otherwise.
00:56:43
Speaker
And he he he said something, can get this right, you have to choose your heart. Living on a budget is hard. Being broke is hard. Like being in debt is hard. so So pick pick pick which one you want. You just remember that yeah I don't think there's any way ah out of it. um But you choose your heart. Right. Being unhealthy, having a lot of medical preventable medical issues. You know, I know that's not everybody's situation um is hard.
00:57:12
Speaker
training and taking care of yourself is hard, right? Like not eating that extra brownie, you know, and, and, you know, I'm, I'm far from a perfect role model, but he had ah a really interesting insight about, about choose your hard because, um, it can be hard now or hard later.
00:57:27
Speaker
And, uh, it's almost always more worth it for it to be hard now than hard later. I love that. And that goes along with, I think I said this on podcast recently, but maybe not.
00:57:39
Speaker
Someone at some point said,
00:57:44
Speaker
easy choices, hard life, hard choices, easy life. And i probably say that to my children far too often, but I think that there's a lot to be said, not just about perspective but and perception, but also that, you know, going for a run is a lot harder than walking or not moving at all.
00:58:10
Speaker
But when you're going for that run or walk or whatever it is, you are striving for some goal, for some outcome.
00:58:22
Speaker
And so just because it's hard does not mean that there's something wrong with it. There might be something wrong with it. Maybe you're running too hard, but you know, broadly there is value in choosing.
00:58:37
Speaker
I don't mean to be sort of, I don't know, hokey or bro science here, but there's value in choosing the harder path. I think it's become a ah component of my decision-making criteria. I have to make sure, you know, and and I think that it's, it's a challenge to even get that far, but, but, um,
00:58:57
Speaker
you know not I'm not choosing this just to suffer, but I'm not shying away from the hard choice because it's hard. um Absolutely. i think I think that's one of the key components of of my decision-making framework.
00:59:10
Speaker
Along with the question, what if the opposite were true? Those two things have been very helpful for me. Jesse, I appreciate you recording with me. It's always nice to talk to you. And so thanks for being here today.
00:59:22
Speaker
Absolutely, Eric. Appreciate you having me on. Thank you.