This webinar introduces WeatherOptics, a weather impact intelligence platform that translates weather forecasts into actionable operational insights. Scott Picariello, CEO of WeatherOptics, demonstrates how their AI-powered system uses proprietary models to predict specific impacts like power outages, flooding, road conditions, and wildfire spread with three-kilometer granularity globally. The platform ingests real-time weather data, non-weather environmental factors, and historical incident data to generate risk scores from 0-10 that update every 15 minutes with forecasts extending seven days. Matt Cronin from Juvari shows how WeatherOptics integrates seamlessly with WebEOC emergency management software, providing automated alerts and visual dashboards for facility monitoring. Key features include the HYPR AI weather model that reduces forecast error by 40%, routing intelligence for shipment optimization, and comprehensive impact modeling covering power outages, floods, road risks, and wildfire spread. The session includes live Q&A addressing data accuracy, global coverage, validation processes, and practical applications for emergency management and operational planning.

Transcript

Welcome everyone to our webinar titled predicting impact before it happens, turning weather data into an operational advantage.

Before we begin, just a few housekeeping items. This session is being recorded and will be shared with all registrants after the event.

All attendees are muted by default to minimize background noise.

Submit questions anytime via the q and a panel. We’ll address them live as time allows and get back to you on any we miss.

If you have technical issues, try refreshing your browser or simply leaving and rejoining the session.

We’ll ask two poll questions during the event today, and we’d love your input on these.

The length of today’s webinar is forty five minutes, and we’ll send a recording and a short survey to all registrants after the event.

Our agenda today, we’ll welcome today’s speakers. We’ll get an introduction to our partner, WeatherOptics.

We’ll get a live demo of the WeatherOptics solution, followed by a live demo of WeatherOptics integration within WebEOC.

We’ll talk about our future webinars, how to learn more, and we’ll wrap up with all of your questions.

Today’s speakers, Scott Pecoreillo is founder and CEO of WeatherOptics.

Matt Cronin is VP of solutions engineering here at Juvare. And I’m Jeff Urkevich, director of partnerships at Juvare.

Before we move on, I’d like to present our first poll question.

Excellent. Please take a moment to help us out with this question. What is your biggest challenge in using weather or environmental data operationally?

And we’ll take about a minute to answer that question.

This is great. Thank you everyone for participating.

Alright. About fifteen more seconds.

Great. Okay. I’ll go ahead and wrap up that poll. Thank you all for your participation.

Excellent. Thank you.

With that, I’ll turn things over to Scott. Scott Picariello, please take it away.

Thank you, Jeff. Hey, everyone. My name is Scott Picariello. I am a cofounder and CEO here at WeatherOptics. Today, I’m excited to take you all through who WeatherOptics is, what we do, and how we’re changing the way organizations and teams understand weather impact.

Weather optics is the leader in a category that we call weather impact intelligence. So not only do we deploy AI weather models that improve the weather forecast accuracy, but we’ve also developed a global impact model that works down to a three kilometer granularity, and that will be the focus of today’s webinar.

Today, we’re trusted by half of Fortune ten enterprises, and we work with a larger variety of clients from local and state government, to health care, to supply chain logistics. We power organizations to take a faster, smarter approach and action before disruptive weather.

What makes WeatherOptics unique as a company is that we don’t actually really forecast focus on weather prediction. We are all about translating weather data into real operational impact. So rather than simply predicting weather conditions like rain, snow, wind, or, you know, alerts, we tell you how that is going to specifically affect your operations.

And for us, the way we do that is through these intelligent impact models that we’ve developed that remove guesswork from traditional weather models. So this is just a a deeper look at how we really differentiate ourselves during specific severe weather events. So, you know, traditional weather prediction might tell you watches, warnings, rainfall, snowfall, more broad information about impacts. With weather optics intelligence, we tell you why you should care about the weather and ultimately what actions you should take all through this sort of impact modeling. So that could be anything from rerouting a shipment or evacuating a neighborhood due to flood risk, changing departure times, not only doing that in real time, but doing that in a predictive sense, hours or days ahead of time.

And the actual engine for developing these highly accurate predictive impact models is is pretty complex. We have to gather, you know, millions of data points to to both train the data and also to actually deploy these models. And it all starts with our AI weather model that we developed about two years ago in house called HYPR. What HYPR does is it takes these global high resolution predictive weather models, blend it blends them together based off of accuracy and performance, and then we ingest real time global observation data.

So that could be radar, that could be satellite data, that could be any surface data that we collect, and then we bias correct the model output based off of these active changes. So a good example would be if weather models were predicting a clear day and a pop up thunderstorm happened. What hyper does is it scans radar and these local observations and will then correct the actual model output to show that thunderstorm. Or maybe a hurricane is predicted to be thirty miles east, but observations show that it’s leaning a little bit west.

HYPR is able to scan and understand that the models are a bit off and will bias correct that output. So over the last two years, we’ve seen really remarkable results from this model, up to a forty percent decrease in actual weather forecast error for things like precipitation and wind gust.

The second most critical step here, and this is really the most important part, is that we ingest contextual environmental non weather data to really bridge the gap between weather and the real world. So we’re analyzing everything from road networks to tree height and density, flood zones, distance from river, river geography. We also look at past storms and how they’ve impacted different locations. That could be storm damage reports, road closures.

We collect data from over forty million connected vehicles. All of this data is super important because it allows us to understand how weather impacts different regions or different locations in different ways. An example we always like to give is one inch of snow in Chicago versus one inch of snow in Atlanta. Same weather, right?

But the impact on what’s actually happening to key operations is entirely different. And so this sort of data layer using the non weather information allows us to capture that. And the result are these highly precise predictive models that can capture how weather will actually affect your operations and assets.

So here’s a look at our core impact models. This is, you know, most of what we offer. There are some on here that are not shown.

Most of these models provide data on a zero to ten risk scale. They update every fifteen minutes. They provide hourly forecasts out to seven days. And as we mentioned earlier, the data has provided a three kilometer granularity on a global level.

So I’m gonna run through these briefly here, and then we’re also going to break down a couple of these in more detail. So we have our power outage index, which predicts the probability of power going out, the number of customers that will potentially lose power, our flood risk index, which predicts where flooding is likely to happen, the severity of that flooding. Road risk index is pinpointing where along a given roadway there might be a risk of danger or closures or delay times. Our life and property index, which tells you how effective weather is going to be for potential issues to property damage or to livelihood.

We have what we call our shipment impact index. It’s really called right route. It will analyze active or upcoming shipments and routes and give you information about how weather is going to affect those. And then we also have a wildfire spread index.

We also have a wildfire conditions index, but these indices basically track active wildfires, and then we’ll predict where they’re going to spread over a twenty four hour period.

So we’re gonna do a bit of a deep dive. We’re gonna start here with the power outage index itself. And as I mentioned, it’s predicting the probability and the expected number of customers to lose power anywhere around the globe, down to that three kilometer granularity.

The way this specific model works, because every model works slightly differently in terms of what we ingest and what it’s using to make these predictions, we’ve trained this model in ten years of actual outage events. That includes hundreds or thousands of storms, you know, data from over fifty utility companies. And then we look at all of that non weather data. So things I mentioned earlier, tree height and density, the type of composition of of certain trees or forests, population density, all of these non weather variables, including environmental ones like topography, elevation, level of exposure to certain conditions, right? Is an area used to higher wind gusts or heavier rainfall?

And then we deploy hyper, that AI weather model, for all of these different weather variables to then make those predictions real time and into the future. What we get is this incredibly accurate model that is able to make these predictions and give you confidence in where outages are expected to happen. This is just one case study that I wanted to highlight. This is from Hurricane Ian in twenty twenty two.

This is zoomed in on parts of Central Florida. What you’re seeing here is a five day out forecast. Those tiles there that are in pink and purple are showing what WeatherOptix was predicting five days ahead of time. So we were predicting widespread fifty to seventy percent of customers would lose power across this area.

Those hot pink areas are actually eighty to ninety percent power loss, and then five days later, once the storm actually made landfall and moved through, we found widespread forty to seventy percent of customers with power loss and some localized areas of of higher power loss up to eighty or ninety percent. So this was a super accurate forecast that we had about five days in advance.

Next, we’ll look at our flood risk index. So again, this is identifying areas that are at high risk of flooding. Not only that, but the severity levels and what to actually expect from an impact lens at a three kilometer resolution.

Similar to the power outage index, we train this on actual flooding events, right? So hundreds or thousands of storms, flash flood events, understanding how local susceptibility changes flood risk, also looking at topography, estimating what kind of rainfall it takes over a certain area to cause flooding, looking at river geography, distance from river, soil moisture, all of these different non weather variables to train and then also make predictions on this model, and again adding in hyper for that predictive weather intelligence. This is a case study from Hurricane Helene in twenty twenty four, which as many of you know devastated many areas, including the Appalachian Mountains.

And so what we’re showing here is a four day forecast where in those dark shades of purple, Weather Optics was highlighting well ahead of time significant or catastrophic flooding in this part of the country, and what actually happened four days later was an incredibly significant event that not only led to significant damage, there were washed out roadways, significant loss of life, and so WeatherOptics in this model was able to identify that days ahead of time.

Next up here, have our road risk index. So this is meant to predict where there are going to be segments of roadway that have higher risk, potential delays, and closures.

This one’s a really interesting one because we’ve actually trained this on live connected vehicles. So we worked with various telematics providers and companies to collect data from over forty million vehicles, understanding how weather is interacting and impacting these vehicles live.

We’re also looking at road classification, curvature, slope of roadway, and that local susceptibility piece in human reaction is super important, right? So how do drivers react in Southern areas versus Northern areas, or even different parts of the globe? All of that is added into this road risk score and index. This is just one case study. We have dozens of case studies in our road risk score and these other indices, but this one is really interesting. This was from I-eighty on Donor Pass in twenty twenty four.

We had a three day shutdown of I-eighty. About five days ahead of time, Weather Optics was predicting a road risk score of eight or higher, which meant a high likelihood of road closures, and you can see here on the top right all of the different predictions. In that bluish purple, have our five day out forecast, in the red, the three day out forecast one days and then six hours ahead of time, and we also marked where incidents and collisions started happening, and when that actual road closure started taking place. So not only were we accurate five days out, but very consistent in understanding that there was a high risk of road closures days in advance. This ended up costing over three hundred million dollars for this closure, and there were multiple accidents that took place as well.

The last one here that we’re gonna do a deep dive on is our routing intelligence product. It is called RightRoute. Essentially, RightRoute does is it takes an origin, a destination, and a departure time, and it creates this analysis of how an active shipment or an upcoming shipment or route is going to be affected by weather. That could be weather adjusted ETAs or how slowed down or delayed we think that shipment’s gonna be.

Our road risk scores included in RightRoute, as well as our other indices, temperature risks, route optimization, all of this is a part of RightRoute. Similar trained off of those forty million connected vehicles, it’s leveraging routing providers as well, so we’re looking at Google, TomTom, Here, Trimble, all of these different routing intelligence providers, to understand whether it’s impact on actual routes, and then we could connect with key third party systems, so ELD, telematics, TMS, routing guidance, and we’re able to make really accurate predictions on how a route or shipment is going to be impacted.

Highlighting this case study, which is interesting, from Maine in twenty twenty four, this is a route from Portland to Bangor, Maine. You can see here basically Google’s ETA predictions for this route in green versus WeatherOptix going out to one hundred and twenty hours. What you’ll notice here is that routing providers themselves inherently don’t account for predictive weather or weather slowdowns, right? So any of these routing providers you might use when planning logistics will show you no delay in a predictive sense from weather, while WeatherOptics three, four days out was already showing significant slowdowns ten, fifteen, twenty percent.

And then you can see in the black dotted line at the very top is the actual observed slowdown on the shipment. WeatherOptix was very close to capturing what actually happened, and this is a one off example, but again, we’ve run this on thousands of routes to gain accuracy and insight into how this product works.

And not only do we have these models, but it’s super easy to make decisions off of them, and to really automate that workflow. So what you’ll see in a second here is that we basically have a four step process. Step one is we create almost like a digital twin of your operations, getting your key assets, routes, facilities, all of that into our system so that we can track it. You can configure your your team as well.

You then are able to set risk thresholds and, you know, different alert configurations. So defining what matters to you and how you want that information conveyed to you, what team you want that information to go to as well. WeatherOptix is always monitoring risk in the background, so we’re checking with our risk models, the real time and predictive impacts, and if it measures up or breaches any of the thresholds that you’ve set. And then last but not least, if it does, we are pushing those insights directly to where you live.

So that could be in the WeatherOptics platform, that could be in WebEOC and Juvari, or that could be, you know, via text, email, anything that you are using, or you’re living in day in and day out, that’s where the data is pushed to.

Awesome, We’ll switch over here now to just a quick live demo of the WeatherOptics platform itself. So let me go ahead and share my screen here.

Alrighty, so this is our Impact Intelligence platform and portal. It’s a web based portal that customers can access. You’ll notice here a couple of important things off the bat. We have two types of forecasts. You can view our nowcast forecast. So this is everything real time or short term that’s happening, weather wise that might impact your operations. And then we have our extended forecast, which goes all the way out to seven days at the hourly level.

We have different types of assets that you can add into our platform, locations, vehicles, and shipments. So our data will go down to the asset level and, you know, really, provide forecasts for the types of assets that you care about. We’ll dive into those in a second. But I want to first start here with our layers, because this is going to show all of the different modules that WeatherOptics offers. So if we go to, map type here first, you’re going to see all of our raw weather data variables powered by hyper. So anything that you guys could want from, you know, rainfall to snowfall, wind gust, wind speed, all of that data WeatherOptics produces and is available, you can scroll through time to see how that changes.

But most importantly, in the context of this conversation, you can also see all of our impact models. So that includes our road conditions index visualized, our business disruption index visualized, and I’ll show it at a global scale as well.

Our flood index visualized here, our power outage index, life and property, and then we have wildfire conditions and wildfire spread. So if we go here to NowCast, I can see our conditions index predicting, you know, where wildfire conditions are happening in the US. And when I go to events, I will show you spread. So you can see any of these, key variables that we have, all of our impact models visually on this map.

You can also see critical events that are happening, so not only the modeled layers, but anything from hurricanes to earthquakes to lightning strikes. We are also collecting data on issues that are happening in real time, so that could be live power outages that are taking place. You can see it by county. You can see weather events at a thirty thousand foot overview.

So looking at winter events or severe weather events and getting a briefing on those events and what’s happening in terms of asset impact, storm damage reports, wildfires, all of these different types of critical events that you can see right here in the platform, including one of our newest releases, which is severecast. So our take on predicting severe thunderstorms. Essentially, are providing updates every sixty to one hundred and twenty seconds, showing you various risks for severe weather. So all of these different layers are available and you can view them with your assets in the platform.

We also have what we call our live road status layers. This is going to show you everything from live traffic cameras. Can zoom in here and see data on live conditions on the roadway. I can see traffic incidents.

I can see closures, truck warnings, all in one singular place. And we also have all of your government alerts here as well. So visually I can see any of these layers really easily and interact with them, and then I can go asset by asset and also see what’s going on. So I can click on a given location and it will show me my specific risks here for a given location on the left lightning, severe thunderstorm event.

There’s some road danger and some other impact indices that are popping up, and then I can see that timeline view. So for any of my hyper local weather variables or any of my risk scores, I can see how risk is going to change over time out to seven days. It gives me a really easy view of being able to see that all in a singular place. I can also look at an impact timeline, I can see all of those risks together in the same place.

I can also see government alerts or weather events that are expected to happen. So that’s at the location level. I will show you at the vehicle level. We connect with your ELDs, so we’ll show you a live look at your drivers and what sort of impacts they will have.

I’ll zoom in here into the Montana area. You can see this driver here has moderate road danger, a risk for tipping or rolling over. So similar locations, but on the vehicle level, I can see those conditions, also recent messages that I’ve sent out to that driver.

And then last but not least on here, we have that right route shipment intelligent piece. So you can see loading in an origin, destination, departure time, and getting all of these insights into how this shipment’s gonna be impacted by weather, delay time, adjusted ETA, road risk, weather conditions, available here in the platform. So this is for a shipment departing on the twenty first here at three pm. You can see where those high impacts are expected to happen.

And you’ll also notice that on many of these options, we throw this yellow flag out. That means there’s a route alternative that we found that might have lower risk or a faster travel speed. So this is the sort of optimization that you can do in our platform.

I’m showing long haul routes here, but you could do this for any amount of distance, right? So even within a city or state or even within a small town, can create routes, maybe an evacuation route that you’re testing or scenario planning, and we will give insights on the safest and fastest way to route around significant weather. So that covers all the different types of assets that you can see in the platform. The last thing that I want to show on here is that you can set up those automated alerts super easily in here so I can go in and set, for example, driver notifications. If I wanted to set, you know, notifications for my drivers every time they entered high road risk or you know tipping risk or wind gusts over forty miles per hour. I can configure that in here really easily. You can see all the alerts that I’ve configured for my drivers that they receive that audibly in a proactive sense.

I can do the same thing for locations or for routes. So these are some, past risk scores that I’ve set up for, risk score alerts that I’ve set up for different locations, but it’s super simple to just do that here in the platform, and set those up in a similar way, craft that alert to how I want it, and then it will be when those thresholds are reached, those messages will be sent out in whatever way you’ve set it. So super easy to configure here in the system.

That just about covers it for the live demo portion of things. You can also save views and different configurations here in system. So, with that, I will stop sharing my screen, and I will pass it back off to Jeff.

Thank you, Scott. That was excellent. Appreciate your presentation.

We are going to go to Matt Cronin in just a minute. Before we do that, I’d like to present our second poll question, and promise it’s the last one. How does your organization currently use predictive weather intelligence in operations?

We’d love your feedback on this. If you don’t mind, it’d be great.

And we’ll take about a minute to, to respond on this one.

Great. Just a few more seconds on this one. Thank you again for your feedback.

Okay, we’ll go ahead and close out that poll.

Thank you. And we’ll turn things over to Matt Cronin. Matt, please take it away.

All right, thanks Jeff. So yeah, we wanted to talk through a little bit how the powerful WeatherOptix data can be fed directly into WebEOC. We have an out of the box facility status board that can be delivered turnkey with this integration.

We can also integrate with other custom boards, and so we’ll walk through those different options, but that will give you a lot of those insights, not all of the insights of course that Scott showed within the WeatherOptix system, but many of the most powerful things and specifically those risk scores. And obviously that allows you to make some good decisions based on the projected risk to your facilities and the impacts to your operations. So let’s go ahead and jump into the demo. I’m going to go ahead and share my screen here versus PowerPoint.

Jeff, can you tell me when you can see my screen? I see you, Matt. Alright. Very good.

So what we’re looking at here is WebEOC Nexus, so the latest version of WebEOC. And with Nexus, have a home page where you can put all of your workflows and maps. And in this case, really this is many of our partner integrations that we have. And so today, obviously, we’re focusing on weather optics, and so we have a number of things to show.

The first one, as I mentioned, is this facility status, weather optics integration. So what this allows me to do is have all my facilities.

I can track things like the status, generator, water power, all of those sort of things.

And then we’ve also integrated with WeatherOptics. So all I have to do is click a record here, and I have a slide out with the details of the projected impacts to the different indexes, those being road conditions, business disruption, flood, power, life property, wildfire spread. And I can click on any of these to turn these off to kind of focus on the ones that I care most about. And I can also hover over and see the specific day and time and specific risk score that we could expect.

We can also see here, we could filter by the facilities. In this case, these are just airports throughout the US.

And I could see which ones are projected to have significant impacts from weather.

I also have an insights tab, which is really handy because this allows me to look at those different indexes broken down. I can sort by them. I can hover over the score at any on any of these areas. And what that will do is give you a really nice description of what you could expect in terms of those impacts to, in this case, Calgary Airport.

And within the next twenty four hours, twenty four to forty eight hours, forty eight to seventy two, and then seventy two hours is kind of the max time that we’re focusing on here. And same thing, can go over here and hover over these different time periods and see what the impacts are going be at different times. And then, of course, you have a dashboard just to kind of summarize our status of our facilities. But this is only one way that well, this is one example.

First of all, this could be tied to other custom workflows, could have similar insights, with other data that you’re monitoring or whether that be personnel or locations or other resources.

You also have the ability to bring a lot of this data in via map services. So as they can bring these things through Esri, some different common JS formats. We had a number of them, and we’re working on now bringing in that weather index layers. So I can see my road index.

I can see my power index. And then as I click on any of these polygons, I can get a description of what that impact is. And I’ll actually see that area that’s going to be impacted by that, you know, as part of that index. So really powerful information.

We also have a version that we’re working on that’s going to have the ability to see this over time. So you have this time slider that you can look at, you know, over a period of time and see how those will adjust over those times. And again, at any point, I could click on these polygons and it would give you the description of what those impacts are going to be based on that period of time. And again, not just one index, but all of the different indexes that WeatherOptics provides.

So really good insights can be available right in WebEOC, and of course, would have access to the full WeatherOptics platform as well. A few other things here before we turn things back to Jeff.

Is one layer that weather optics provides a number of different layers. This one with weather events, I really I really like. It’s also I think we also have lightning in here, you can see where lightning has occurred. But if you click on these polygons, a similar type of information, but it’s more of a briefing, right?

So we have severe weather. This is what the impact is. This is where it’s going to be impacted. There’s various stats in terms of wind speed, rain, flooding, you know, just all the sort of information you would want and need is available through the pop up here through this integration, again, both with workflows and as part of this map based integration where we’re pulling this data in.

And then finally, just some other layers that and these aren’t all of them, of course, but we have, you know, earthquakes and lightning and power outage information, you know, volcanoes if you want it, fire, and then any of these you can click on the particular icon in the area that you care about, and then see, alright, this is a wind gust, fifty five mile per hour. This is the National Weather Service office where that’s closest to this location, and when this information was, you know, last updated. And that’s just that’s just an example. But again, you can create any number of maps, you can overlay your WebEOC data with this, and you can, you know, have access to a large number of map layers that WeatherOptics provides, or as mentioned here, we can have an integration where we can basically talk to WeatherOptics every, five or ten minutes, whatever frequency we define, and we can get some insights for your facilities inventory and resources.

But let me go ahead and turn it back to you, Jeff.

Thanks, Matt. Excellent. Thanks for that presentation. Before let me just share my screen again.

Okay, I encourage you to please hang around for questions. We are we have a lot of questions, so please stay on. If you wanna learn more about anything we’ve talked about today, you can reach out to me directly. I’m Jeff Berkovich, and my email is there.

Same with Scott Picarillo at Weather Optics. You can request a demo through our website, or you can learn about any one of our partners on our website also. So duvara dot com slash partners. I’d also like to mention our Elevate conference.

So Elevate twenty six is happening in May, May twelfth through the fifteenth. Great opportunity to to collaborate with other people, to to learn more about our software, more about our products, and what’s coming up with the with Jovari and WebEOC. It’s a four day event. I really encourage you to join us.

It’s free, and it’s open to everyone. You don’t have to necessarily be a a customer of Gevari. You can you can come and join us. Sign up for that at elevate dot gevari dot com, and we hope to see you there.

We also have a few, future webinars. So SENSE NET wildfire detection and ice eye flood detection, both of these are coming up here over the summer, and we’ll be back in touch with, some some dates as they, as they, get planned.

Okay. As I said, we have a lot of questions, so I’m gonna go ahead and throw these out to to the team. Just one minute.

Find our questions. Q and a. Here we go.

Okay. And thank you all for the questions. Let me start at the top.

Do you ever run into any issues with bad or incomplete data? And it sounds like maybe a question for Scott.

Yeah, it’s a great question. For our risk scores and any of our predictive, hyper local weather layers, we do not run into those types of issues.

For certain things like some of the event data, there’s occasions where at a global level level, one of our sources could be incomplete. But for anything that we produce in house at Weather Optics, we don’t run into any sort of incomplete or bad data issues. We’re able to provide that globally after seven days.

Thanks. It looks like a follow on to that question. Can it pick up on chain reactions like one failure causing another?

Potentially, yes. As we collect real time reports of like road closures or other issues that could be further exacerbating issues, our risk scores can tune up to those and understand that you know, certain things are taking place and adjust our ratings based off of that live information.

Excellent. Now this may be for Scott and Matt. I’m not I’m not sure, but I’ll ask the question. Can this data be used to help emergency managers monitor evacuees evacuating road maps during a wildfire? In other words, can this data show concerns of evacuation routes during a wildfire?

From from our end, yes. Right? So for our routing guidance that I showed earlier, the the right route intelligence, it will include if there’s an active wildfire that could be impacting a given route. So you can load in evacuation routes or roadways into our system, and we will show if any of those are at risk of wildfire spread.

So it can help during those times, and that can also be shown in the Juvari WebEOC platform.

Great next question, does the wildfire layer include atmospheric conditions and plumb modeling on those conditions?

It does include atmospheric conditions in terms of I think plume modeling maybe is is what was perhaps meant there. We have some ability to do modeling on air and issues that could come from a wildfire, but this is something that we’re working on enhancing. We’ll actually have a plume modeling product here before the end of the year, that will additionally cover, I think, some of that information.

Great. Does WeatherOptics cover provinces in Canada?

Absolutely. We we cover global data, including Canada. We have great data for, Canada and anywhere in North America or around the globe.

Okay. Does the overall what is the overall accuracy of Weather optics compared to what actually happened?

It’s a great question. We are consistently collecting real time incident data and then comparing it to our risk models. Right? So for the road index, for example, we’re making these predictions every fifteen minutes or every hour, and then we look back retrospectively at given events and see how close are we to predicting an actual incident, or are we over predicting or under predicting?

So we have a lot of metrics and statistics by individual risk scores. So the overall how accurate is WeatherOptics is more challenging. We can provide information by risk score or by weather variable. I’d mentioned earlier improving or decreasing forecast error for weather by about forty percent compared to some of the short range governmental models, and we can also provide metrics for those risk scores as well.

And we track those in the platform itself, you can see for given events exactly how we predicted them.

Excellent. Speaking of predictions and history, is there any way to look retroactively at the historic power outage data with this product?

Yes, there’s potential to be able to look back at both what we predicted or simulated and estimated with our power outage model, and also because we’re collecting live outage reports, it may also be possible via our API to see actual historical outages that have taken place.

Excellent. The questions are coming in faster than I can keep up. Can this system be applied for any country?

Yes, it can be. This data is global, right? So even when we’re ingesting non weather data that is informing how impact might unfold, even if we don’t have that coverage in certain parts of the world, we are still simulating and estimating based off of whatever knowledge we can gather from different parts of the world, where outages are likely to take place or where road danger is high or where flooding risk is is likely to happen. So it works anywhere, any country around the globe, and has also been tested and validated in other countries as well.

Excellent. Does plume modeling include hazmat releases?

It will include hazmat releases. So this is a product that’s in development right now. The current wildfire product itself has some basic kind of plume modeling and air quality information.

The the actual plume modeling product itself that’s released later this year will will include more of that.

Great, you’ve this may be too much for this call, but I’ll go ahead and ask the question. Can you speak to your validation process for the property loss module?

Yes. It’s it’s a great question. So as I I mentioned a little bit earlier, it’s similar to how we’re validating the other indices where we are collecting damage reports and loss reports from around North America, around the globe, and then comparing that to what we were predicting on that zero to ten property risk score.

So we’re consistently validating how accurate we were, and then also improving the model. These models improve over time because of that data collection. So if we were off during a hurricane or a given event, our models are learning from that information in real time as well and improving with each event that happens. So hopefully that gives you a bit of a sneak peek into the process.

Okay. For your wildfire spread index, do you have Canadian fuel types? I’m not sure what that is, but I’m sure you do, Scott.

Yes, but would would want to learn more about what specifically is is being asked here, but I I believe we we account for different fuel types.

Okay.

Are there any other things that I want to Sorry, there was one question I want to make sure we covered because I didn’t cover it sufficiently, and the question was, can the data be printed or emailed?

And it absolutely can be printed, it can also be emailed. Scott showed how the data can be emailed from WeatherOptics, and WOC has similar automated emailing capabilities. So if a facility is going to be impacted, we can automatically send an email to the folks that need it to let them know. It could also be text message or voice call as well, but that automated alerting is definitely something that can be done.

Excellent. Thanks, Matt. Do you have a capability to do rapid damage assessment? If so, which hazards?

We have some capabilities to do damage assessment. I would need to understand the level of detail that you’d be interested in. We are collecting damage reports consistently. So as reports come in from news or from National Weather Service or for government, we are analyzing and collecting those reports and putting them in the platform. They would include all types of weather hazards or damage.

Okay. How accurate is your flash flood prediction and modeling? Have you experienced any issues with flash flood predictions?

This is one of the areas that we focus on the most with our flood index. So we can provide information on how accurate we are overall. We also are able to look at individual events. So if the folks listening have certain events that they care about, an event that happened three years ago in this part of the world, we can retrospectively look back and show you what we were predicting four days out, five days out, and actually give you accuracy for those individual events. So just something to know. It’s a common use case with our system.

Okay. Still a few more questions. We’re gonna run out of time soon, but we’ll go ahead and answer as many as we can. Is your predictive system ingesting information from nonauthoritative sources? For example, traditional media and social media posts.

We have some of that, but everything is is vetted in our system, so there is confidence and accuracy types of scores assigned to everything. So if it is coming from a non authoritative source, that is also being accounted for. Most of them are from authoritative sources that we’re using in order to ingest that information.

Okay. Can you give an example of how the model has been used for planning purposes with file with wildfires?

Sure. So for planning purposes, if there is an active fire, what our spread index does is it will give a probability how likely that fire is to spread to the next nearby point, right? And that can then go down to the route level, the vehicle level, the location level. So we’ve seen times, for example, like we were talking about earlier, where there might be certain evacuation routes that are preferred by a state or a local government, and our RightRoute software will use that wildfire spread index to say, Hey, there is a high likelihood in the next twenty four hours that that fire is going to spread over that route, and so you should choose another option. So I think that’s a great tangible example of where the index can be used and then where the actual asset level information can also be leveraged.

Excellent. Okay. We’ll take one more. We’re actually over time. Are there any AI components that run-in the background of the platform?

I would need to understand more about what background of the platform kind of means.

We are deploying AI to create and generate these forecasts. There isn’t really AI that’s actively, I guess, in the background beyond the the models themselves that we’ve that we’ve created.

Okay, I’m going squeeze one more in, Scott, because this is a good one. Are there datasets strictly using publicly available data or are there proprietary data used for the event data and impacts?

The event data and impacts is a combination. It uses proprietary data as well. When we’re creating those critical events, we are looking at our risk scores in order to generate those. So it’s a combination of public and proprietary.

The city states used strictly publicly available, or are they any proprietary for the event? Yeah, so it’s a lot of proprietary data. The risk scores are creating and generating those events.

Excellent. Okay, everyone, if we didn’t get to your questions, we we, can respond directly. I’ll reach out to you via email. We’ll respond to your questions. Thank you again for joining, and have a great day.