You are listening to the HumAIn Podcast. HumAIn is your first look at the startups and industry titans that are leading and disrupting artificial intelligence, data science, future of work and developer education. I am your host, David Yakobovitch, and you are listening to HumAIn. If you like this episode, remember to subscribe and leave a review. Now onto the show.
Welcome back to the HumAIn Podcast, today on the show we have Jared Goldberg¹, the Head of Data Science at WeatherOptics. Now for many of our listeners on the show we have had before thought leaders, who have worked at IBM involved with the weather channel and many different startups in the industry. And one of the big things we’ve known of 2020 is that one of the largest acquisitions to date has been Dark Sky.
It is the famed and loved app on the Android and Apple app stores. And I have been a Dark Sky user. Of course, they were not really talking about Dark Sky, but we are talking about how the weather space is heating up. And so today Jared will be our guest and our expert. So Jared, thank you for joining us on the HumAIn Podcast².
Thanks so much, David. I am excited to be here.
Yeah, thank you. I wanted to hear a little bit about yourself as a starting point. You know, you are working on a startup on WeatherOptics, and you’ve got to play with Dark Sky’s data in the past, you have your own data, your own APIs, what is going on, let our listeners know.
Sure. So weather optics is what we do. And WeatherOptics³ started as a weather blog way back when, from our founder and CEO, Scott Pecoriello, he was a weather nut growing up. And he had this blog, it was on Facebook and social media. And his snowfall accuracies were crazy good. He would put out these maps of where the snow was falling the most and where the snow was falling the least for these storms.
And he was nailing it. And he got a very big following doing this and his father is an entrepreneur and the business spirit is in his blood. And soon, he turned it into his business. And in November of 2017, he reached out to me to bring the business into this tech side of things and into the data side of things. And we have just been gaining momentum since then.
Now speaking of snowfall, many of us know that in mid-May 2020, there was record snowfall in the Northeast. There were talks about in Boston, New England, getting up to a foot of snow. And we’re looking back on that now, and I cannot believe how the climate continues to shift.
I mean, today in the episode, we are going to talk a lot about how WeatherOptics is doing, how COVID-19 has been impacting climate, both negatively and positively and where trends are going. But it’s so interesting to see how from snowfall back then to snowfall today, weather is everywhere. Could you share with us a little bit, Jared, about the history of quantitative forecasting and why you are so interested in data science?
Sure. That’s definitely a good point. And these last couple winters have been weird to say the least. And I’d like to hope that people are realizing that the weather and the climate is changing and we think it’s never been more important to understand those trends. So I’ll, preface by saying that I’m certainly, I’m not a meteorologist, I’m a data scientist that works with weather data.
However, I had our meteorologist Josh, give me a really good rundown so I can present some of that here. Quantitative forecasting started in the 1920s and it was Norwegians that took the reins on this stuff. And really the head of modern forecasting is a guy by the name Lewis Richardson. And he started producing atmospheric forecasts by hand back in the 1920s.
A lot of this was physics based. So, there was not a lot of precipitation. In fact, they were not doing any sort of forecasting of precipitation. It was purely things like thermal energy and these really primitive equations building off of physics that allowed them to do this first stab at forecasting the weather and it was the first successful stab of forecasting the weather.
And soon it was widely adapted and he needed more compute power, so to speak. So he got more people in his room to help him crunch these numbers by hand. So it’s interesting in terms of connecting to computing as a whole, whether it was one of the first reasons why people even use computers was to do these sorts of calculations.
And that leads into the 1950s where supercomputers were making these calculations and would take up a lot of space in a room, but things were getting better. So they were starting to forecast things like temperature and they were gauging the atmosphere using a computer.
Things were very slow and an interesting tidbit from the 1950s, is that the first 24 hour forecast took 24 hours to produce and as you can imagine not the most efficient method.
I like to think that WeatherOptics forecasts happen a little quicker than that, which they do, but that’s where it started and had a lot of applications. So organizations like the Air Force used this information to make decisions. And even back then, people knew that the weather had a big impact and so people were working towards predicting it. The next stage of this process was the 1970s where computers became more and more regular. People were used to working with computers then came things like satellites, which give more data about the atmosphere.
And they were making 2–3 day forecasts and using other tools and things were getting more sophisticated, similar to most other technological trajectories like cell phones or machine learning algorithms.
They just kept building on what they knew and making things better and the modern era of forecasting started in the 1980s and that’s where we had global forecasting models based on a more complex system of observations, but still building off these physics concepts that were used originally.
And what’s happened since the 1980s is that things have just gotten more complex. These models have gotten better. And now there’s this whole wild system that no one really realizes is happening where you have all of these different inputs from all these weather gauges, like airplanes and satellites, and they’re all amalgamated and interpolated into these models.
But think about how far out these models go, so as things got more complex, the output of one model would feed into the output of another model and there’d be missing data, right? I mean, we have more data points in some parts of the globe than other parts of the globe. So we’ve had to fill in the gaps and do all sorts of crazy technology to now get us to a point where we have this global modeling framework and we can accurately predict the weather a week in advance to the hourly level. This is what WeatherOptics does.
So again, I’m not a meteorologist, I’m not the expert on these modern weather models. Maybe one day we’ll have Josh, our meteorologist join us here and really dive into it. But it’s really a crazy incomprehensible system that actually allows WeatherOptics to do what we do.
It’s so fascinating to hear how the entire weather industry, this quantitative forecasting has evolved in the last 100 years. I mean, when we accelerate from the 1920s, which was by hand, without machine into 1950s with the machines, I think the 1950s was such a critical time era in the development of many technologies.
I mean, last year there was a significant Hollywood blockbuster that was highlighting hidden figures about NASA black women who were working in tech, who were building these first computer machines we’ve seen with Alan Turing and many other systems that the 1940s through 60s led and that compute acceleration sounds very similar for quantitative forecasting.
And then even the 1980s, that modern era, I mean, when we’re thinking about data science and machine learning, we even see that Stanford’s labs and Carnegie Mellon really accelerated in the 80s where we were exploring self-driving and novel technologies that only today are being practical.
And it’s fascinating to see that Industry agnostic, there was the consistent patterns about the acceleration of digitalization and compute. And that’s leading us to where we are today, with what you’re working on at WeatherOptics. I mean, we all know whether we’re here in the United States or globally listening into the show, the weather impacts us all.
And every year there are natural disasters. Of course, this year, one of the natural, not so natural disasters is COVID. But beyond that, you know, real natural, like earth based natural disasters include things like hurricanes and tornadoes. Earlier this year in Mississippi, there was the F5 Tornado, a mile to two miles wide. And a lot of the news was saying, this is unprecedented, that the weather is just getting more and more severe. What’s your take in general on natural disasters and how weather impacts us as society.
Sure. Anyone who’s ever existed and has senses knows about the weather. We feel temperature. We play in snow, we cover ourselves from rain. So to some extent, everyone has this inherent understanding that the weather changes our behavior. When it’s raining, we bring our umbrella, we put on our rain boots and so forth. So there’s these different impacts that weather has and it’s on these different scales as well.
So as you mentioned, on one hand, we have these big, intense weather events that can cause severe disruption that everyone’s familiar with. Things like hurricanes, as you mentioned, tornadoes are the flashy part of weather forecasting. You have people that get in cars and chase tornadoes and try to understand them and deploy different tools to help better understand their formation and how they’re going to develop and where they’re going to develop.
And a lot of people get really enamored by that. And of course these big disasters have a massive effect on society. I mean, we cannot ignore that. So obviously things like Hurricane Katrina had over a 100 billion in terms of economic impact. Many people lost their jobs. It really changed the United States, similar to what COVID is currently doing to us as well.
So those cannot be ignored and WeatherOptics data certainly picks up on those sorts of things. However, we need to keep in mind that our business and where we really understand the weather better is the short term weather events. So hurricane certainly count as that, but it is not every week that we have hurricanes that happen across the United States.
So the other aspect in terms of weather impact are the short scale weather impacts. Like if it snows a couple inches in Seattle, it is going to have a very big and disruptive effect on that area versus three inches of snow in Boston, will not have that same effect. So it is these nuances that WeatherOptics really tries to understand.
And that is where we think our data is very valuable. So for example, we just talked about how Katrina caused billions of dollars in losses. However, if you were to look at the data from that three inch snow storm in Seattle, you would see that the day that it snowed those 3–5 inches, people spent less money at somewhere like Walmart or people went out to the restaurants less that day because they did not want to drive through the snow.
And it is these sorts of impacts that obviously are not as flashy as something like a hurricane or a tornado, but we feel understanding how weather impacts daily life at these smaller scales and these less major events actually can save people and companies a lot of money and can really improve their processes. So that is where we have focused and that is where we feel we are adding value to a lot of companies that we work with.
It’s so fascinating to think about the short scale, because the short scale really is the long scale. It’s about what’s happening beyond the actual event. You know, this event called a hurricane, a tornado, snow. It only lasts for a short period of time, but what are these extended ramifications? So we look at COVID that is something near and dear to everyone’s life.
Now that the extent of the ramifications is, you know, a month, 15 months, a year, three years, both economically and environmentally we are not sure what that looks like, but it is so interesting that you are making this tangible to weather because you are right.
Three days of snow or a day of snow well, people are not going to be eating out at restaurants right there, because of this disruption to human behavior. And there is an economic consequence there. Now offline, we got to chat more about the business model and what you are working on at weather optics and industry by industry.
I know you are very focused on the enterprise space. That includes, you know, industries like logistics and finance and emergency preparedness. And these are some of the big topics of the hour or the year as a result of COVID. Can you shed us some light on what you are exploring in these industries?
Sure. So as everyone knows, coronavirus has put a huge damper on everyday life. And while some industries have been excluded or cut off, and obviously a lot of people are losing jobs, there are other industries that we are leaning on much heavier. And one of these industries is logistics. So people are ordering online more than ever and that Amazon prime package does not magically appear on your doorstep. It gets from point A to point B. And this historic snowstorm that happened in May was quite disruptive for the logistics industry.
And one major application of our weather data is building useful ways to understand, not just that it is going to rain or I guess in this case, it is not just going to snow in May, in these certain areas. It is to say, if you are driving through the snow, that is happening, how is that going to affect your route?
So one of the big jumps that WeatherOptics takes in the logistics space is not just to supply that raw weather data. We are really taking the onus to give you that signal or give you that understanding of how it is going to impact your fleet. Another industry that we know has taken a huge hit from coronavirus is the financial industry.
The markets are really wanky right now and they took a huge hit and similar lead to what I discussed with that Seattle example, where people are spending less money in certain industries, depending on the weather. That happens every day. Now we will not see those signals as much now, obviously because no one is going out to restaurants.
So if it were to snow, we would not see that effect. But if we take ourselves outside of coronavirus for a second and imagine that this was the same time last year, people who invest in these restaurant companies want to understand their financial statements. They want to understand their performance.
And we are aware that weather has an impact on sales. And if we can help companies understand how weather is impacting their sales, we can help those companies in their own, right that are managing workforces and managing hours and deciding when to stay open and things like that.
And we can also help people who are investing in those companies to understand those trends and make sure they have a true understanding of the performance and what is causing this financial performance. So we consider weather data, a viable source of alternative data in terms of quantitative investing and things like that. We think these weather signals, so to speak can help explain variations in other datasets that help us understand the market.
You know, traditionally when we are thinking about these industries like logistics and more specifically finance such as myself, my background is finance. I come from actuarial science and back office data support for different banks and institutions.
And we are always talking about chasing alpha or seeking alpha where alphas lift or increase that trend. And I’m putting together the fundamentals right now that for weather this could be not necessarily chasing alpha, but chasing the signal. And these are the signals that you are trying to find out and discover through actionable data. So let us get more specific into that data on that granularity. What does WeatherOptics do today to help quantify impact?
Sure. So the bread and butter of our business and what we are building around is the idea that the lay person, the data scientists working at any of these affected companies do not want to grapple with the weather data itself. They do not care about humidity. They do not care about dewpoint. They probably do not care about temperature as its own data point.
So what we are trying to do is take those raw weather data points and turn them into something understandable for people who need to use that data. And the way we are doing that is using what we coined as these “Impact Indices”.
So we have a combination of meteorological expertise, as well as machine learning. And we are combining these things and essentially what we are doing behind the scenes is modeling how a meteorologist would say how bad this storm is going to be. So in an ideal world, I would have Josh, our meteorologist, look at every single place in the United States every single day and tell us for example, Boston is going to get hit badly by this snow storm, Pennsylvania, not so much.
However, there is one Josh and there are millions of data points every hour. So we are essentially building these algorithms to mimic that decision-making and what we give to these businesses or these impact indices that say on a scale of 1–10, what is the power outage index based on the raw weather data?
On a scale of 1–10, how bad are the roads based on this weather data? And we have put a lot of thought into these indices and we do not just use weather data. We also use non-weather data because that is very important to predict these things. So as a concrete example, two different places that are identical, you can imagine them even close to each other in the same town. They might get the same amount of rain. And typically someone might say, for instance, it rained 3 inches. Both of these places have the same flood threat.
However, we have dug below that detail and we have information on how close are each of those towns to the nearest major source of water, because if one of those towns is close to the Charles river, for instance, then it’s flood threat should go up based on that same amount of rain.
So we have really covered all of our bases and we have been very thorough to truly understand how the raw weather data paired with these non weather variables, add up to these actual impacts and we feel by delivering impacts as opposed to raw weather data, we are going to allow businesses to make impactful decisions that way they do not have to wrestle with the data itself.
It’s so fascinating that you are working all these indices because as data scientists, we often think of these as benchmarks or machine learning that we are making predictions on how healthy or robust a certain factor or variables can be. And, you know, looking through the indices that WeatherOptics has, you know, some of them include like business disruption, traffic speed reduction, power outages, road conditions, temperature, flood threats, like you mentioned and precipitation.
Some of these I think are extremely timely and relevant. I mean, we think about business disruption today, that is of course what many people are thinking about. But business disruption of course, traditionally is from the weather, right? Not from COVID, but even what is so interesting is traffic speed reductions. We had the big snow storm in the Northeast, then you know, is there planning being put in place so that people do not crash so that cities are developed safer? And that last mile delivery of logistics and finance is being properly handled.
How about even power outages? You know, I think these are such interesting use cases and I want to specifically dive deeper into both the traffic speed reduction and the power outage indices. With some use cases I think are really fascinating. The traffic speed, you know, myself being a Floridian who is now a New Yorker by association of being here in the Northeast.
I see all the time, cars that speed through the lanes in New York City and especially around COVID one thing that has become clear is that everyone from New Jersey is coming to the city with their Lamborghinis speeding down the roads as if they are in the Malibu highway or you know, Justin Bieber just cruising in South beach for a party.
And it is so challenging to see this because although this is not weather related, this is still an indication of traffic speed and how this is changing over time. And I think what is most fascinating is how different events in life cause an impact to the roads and to our cities and whether it is COVID related or a new technology related, we are going to see that.
And the big technology that has been taking over conferences like CES and, and others is self-driving cars. Becoming closer to reality, it is a work in process. In 2020, the CES conference showed that Mercedes-Benz now can have a BMW motorcycle to self balance itself. So when we are watching movies and shows like Westworld, these bicycles and motorcycles can do it.
Now, we are getting there. But the question I have for you reigning this all back in Jared is, you know, thinking of self-driving cars with weather. I mean, weather based conditions have so many signals and indices, how are you thinking about the future of self driving cars with weather and how that may impact our cities?
Sure. Obviously, this technology is really taking off and you have a ton of these companies popping up trying to be the Vanguard in terms of developing the best, the safest self-driving car and naturally something they all need to keep in mind when building this technology is the safety aspect of things.
And specifically, just like your normal car has an antilock braking system that tells when you are sliding on the roads, we expect that these self-driving cars will need to have an even better safeguard against these road conditions that could be disruptive to normal driving.
So something that we imagine weather data could help with is giving not only a heads up to the car itself to anticipate those conditions based on where it is and the direction that it is traveling to perhaps know in advance that it is driving into a storm and to prepare those systems correctly and ensure that the safety of the passenger will be upheld.
But also real time, having that understanding because these conditions can change. And again, if they were to use a traditional dataset that just uses typical raw weather data into these readings, they really might not understand how dangerous the roads are. For example, It could be raining and the car might think that it is raining and so it does not do much to actually prepare for any sort of dangerous conditions.
But let us say the temperature was hovering just above freezing, in which case that can become super dangerous because that rain could freeze obviously and turn the roads Incredibly dangerous in a very quick amount of time.
And it is those sorts of interactions between variables that we feel our impact indices would allow people to have the upper hand to understand that just because it is raining does not mean that the roads are not going to be dangerous. And perhaps these cars, these very smart and intelligent cars should know the level of danger and how prepared they need to be in order to uphold the safety of the people using them.
I definitely hope that as the world, we are going to be moving into where these sensors are here and they are here to improve conditions. And ideally those conditions are going to of course, be monitored with ethical AI. I mean, I only think back to the new show on Amazon, which talks about where there are self-driving cars everywhere and spoiler alert, the main character dies in episode 1 in a self-driving car.
So it gets a very heated, contentious show. But it is interesting because episode 1 was prioritizing. Do you prioritize the human or do you on the street, right? Or do you prioritize the human inside the vehicle and what will that look like?
With weather events or supply chain events, it’s super fascinating. Let us segue a little bit to one of your other indices, the power outage index. This one is really fascinating because many of our listeners are across dense urban areas, especially like California and Silicon Valley. And many of us recall in 2019, there was the big California rolling blackouts from PG&E where their electric grid got overloaded as a result of wildfires that were unprecedented because of wind.
There was wind upwards of 60–70 miles per hour, unprecedented at the time in California for an extended period of time, which caused some significant difficulties with their infrastructure, which led to these wildfires that led to, you know, the trickle down effect of all this devastation.
But thinking about power, there are many reasons power may go out because of these unprecedented wildfires, storms, or could be other things such as excess oil capacity as a result of COVID. And where does energy go? So, share with us some of your take on the power outage index.
Sure. So you bring up a really good point that power outages can be in terms of how weather affects humans on a day-to-day level. I remember back during Hurricane Irene here in Massachusetts, my father had just gone to the grocery store and he happens to be a big fan of Ben and Jerry’s ice cream.
And they happened to have a sale on Ben and Jerry’s ice cream that day. So he went out and bought at least five or six pints of this stuff stocking up as the freezer is always full of it. And lo and behold, over that week, Hurricane Irene came and we lost power for an extended period and that ice cream had to go.
Now that is a lighthearted example, but on a more real note, as you mentioned, these weather events can have a massive impact across large areas. And weather is the main reason why power goes out. The grid is so powerful these days that I don’t really know the last time there was the New York blackout. But there have not been that many but just random outages typically caused by, as you mentioned, wildfires or the wind or other factors.
So that is why we have really been focusing on the power outage index. And in discussing these California outages, this would be the perfect use case where if the emergency management companies and government groups that were preparing for these things, if they had a really accurate forecast of what was going to happen in the future, based on the weather, then they could have had a better response.
Obviously their response in terms of some of these preemptive outages and things like that was very controversial and made people pretty upset. So if they had a better plan going forward, and if they would have been notified about the potential of what was going to happen beforehand, they could have planned better.
And that is just a prime example where having an understanding of weather conditions could have made a big change in terms of what actually happened and what ended up happening. So it’s another really important use case for us and it’s one where, where we are talking with people who understand these things and trying to understand what causes these power outages to give the best signal.
Now speaking of controversial signals, you mentioned the story about Hurricane Irene and where the Ben and Jerry’s ice cream had to go. So, here is the big controversy: By to go, do you mean that your family binged watched Netflix and ate all the Ben and Jerry’s ice cream or it got tossed down the garbage disposal?
It was a combination of the two. My father was determined not to let all of it go to waste. So I do believe for the first couple of days that we were out of power, he opened the freezer, maybe against everyone else’s, wishes and dug in as much as we could. On the other hand, I also remember pouring out Gloopy watery ice cream. So it was a battle that was taken on and a battle that was eventually lost.
Oh man. Well, you know, that is of course one that is lighthearted on controversy, but we look at many different scenarios on what can be controversy, because similar to the stories that you shared about like, if PG&E had given more notice, for instance by announcing, ‘In the next month we will start progressively putting these outages’, people could plan their lives around that.
Similar to COVID again, the big thing going on in the world right now, if we have more people know about this, we could have bought masks and hand sanitizer or loaded up in toilet paper without overwhelming the entire industry. But you know, we are now where we are.
I think the supply chains are being disrupted in certain capacities. We are seeing with the meat industry, where industries have leveled off and other companies are adjusting what their just-in-time inventories look like. But, you know, I think the common thread we are seeing whether with these Weather incidences or even with COVID is mitigating loss.
Loss is not something anyone wants to experience, but we have to plan for it. Recently on HumAIn, we featured the CEO of ElectrifAi which is a company in New Jersey where they talked about Contract AI. As a result of everything going on today with COVID, how are you preparing for what could be lost? But, you know, loss is different than many avenues. And then WeatherOptics, Jared, I know you guys are also helping to plan for mitigating loss. What are some of the things that you have discovered as your team has been building your API and your insights portal?
Sure. Mitigating loss is massive and a good way to frame it is that on your typical day, the weather should be good. It does not snow in New York every single day. Now, if you are in the mountains or somewhere with a very specific climate, obviously this does not apply to you, but for most people across the United States, bad weather can be thought of as an anomaly. It does not happen all the time.
So the idea of WeatherOptics and why weather needs to be included in people’s, you know, their understanding of their business is that we have a way, even though it’s an anomaly. And there is certainly randomness, bad weather by definition can be thought of as a random variable with a certain probability of occurring.
However, it is one that we have a grasp that could happen. So the whole idea of our company is these impact indices and all of our forecasts allow these companies to have the heads up to say, we think something disruptive is going to happen. So you should change your behavior in order to mitigate loss.
Now, in some cases you could make an argument that in the financial industry, knowing bad weather could help certain industries. So that is a very specific example where understanding the weather can help you make gains. If you know that, for example, Home Depot is going to experience a spike in sales before hurricane. But in general, a big thing that we are doing is for example, in the logistics space, is telling people to route around the weather. So mitigating loss is huge as we have mentioned, the economic impacts of bad weather are vast.
And so once a company has identified that they would like the heads up about this bad weather, and they would like to understand how weather is going to impact their day to day operations, the whole idea is we want to deliver that information in a format that makes the most sense.
As I mentioned, the whole point here, and what we are trying to accomplish is we do not want other data people having to pull their hair out, playing with the weather data and understanding the weather data.
So we are trying to serve it on the right plate for each of the companies that we work with. And so a good example of that as you mentioned, is the insight portal versus the API. So our insight portal is for more of the non-technical audience. And this is for individuals who perhaps are managing a certain geographic area.
So people that have been interested in this include these emergency management companies that need to understand how their city is about to be impacted at a high level. They do not need the data itself. They just need to view in a visual really easily digestible format what the weather is going to be over the next 7 days. What are the road impacts going to look like in my geographic area? What is the power outage index looking like? They can see the raw weather data if they want. So the insight portal is our attempt at the most user-friendly nontechnical delivery of these same insights.
On the other hand, we are certainly a tech company. We work in code. And we know that there are other companies that are more technologically advanced and are going to want the raw data itself to work into their machine learning models. And our most technical offerings you could argue are our APIs, which are delivering the raw weather data itself, such that we give you those impacts very granularly.
And then your data science team would get a chance to play around with it and use it in the way that is best for them. And lastly, we know that it is not just two extremes here. We understand technicality so to speak is a spectrum. You have people, individuals that really need the visual, just the portal and then the other extreme would be people who are writing their own code.
And then you have that middle ground of people that maybe are working in Excel and are definitely data savvy to some extent, but do not necessarily work in code. And so, the other thing that we are building is this middle ground to deliver things like Excel templates that have this weather data aggregated up.
So it is definitely still tabular data and it is definitely granular and giving what they need, but they do not need to write code to access it. So the whole idea is we want to let people know about the weather that is going to happen so that they can mitigate loss and people have different needs in terms of how they access that information. So we are trying to cover our bases and make it as user-friendly as possible for all companies that work with us.
I think that is fantastic. Even going so far have these customized solutions with Excel templates, because when you think of a lot of the small mom and pop or local weather stations and channels out there, they may not have these huge systems with large budgets.
But doing a lot of traditional work in software like Excel, I know having gotten my start in actuarial sciences that even a lot of advanced modeling is still done in Excel with different cubes and databases today. Of course, there is the correlary that you can build and scale to now power BI and all these Microsoft apps that have been shared recently with all the new updates from the store in the last couple of months. But it is great to see between the API, the insight portal and those custom solutions.
There are a lot of opportunities in the weather space, especially at WeatherOptics. Beyond looking at the near-term though, we want to step back at a macro level because the future is so uncertain and some things are more certain than others, but as we all know, 2020 is a year no one seemingly could have predicted except Bill Gates, but looking at it, you know, there are a few uncertain things I’d love to dive in deep into.
One to start out is about climate change. And the second is COVID, but before COVID, let’s get into the more relevant topic of climate change. It seems that every couple months or a couple of weeks, there are new things coming out, you know, one of the big stories we’ve seen this year is about these swarms of locusts.
And normally locusts are flying creatures that actually help out potentially with different pests and help move crops and animals and help with normal migrations. But with this year, there have been these huge swarms of locusts in Africa and Asia, which have been devastating crops from farmers. I mean, is this climate change or what is going on here?
Sure. So it’s a common theme when discussing climate change in terms of the difference between climate and weather. And it’s really important to understand those differences and how each of them has different ramifications.
But one thing that is clear is that, weather is related to climate, obviously, and that’s why they’re mixed up a lot. And once you start seeing more incidences of certain types of weather over a long period, well, then that’s an indication that the climate is changing. And so what we’re going to start seeing is these sort of weird events potentially happening more often, not a lot of snow in New York city, the last couple of winters and now we have a major snow storm in May, right? That’s a little weird. As you mentioned, these locusts typically stay under the radar, so to speak, and maybe these major changes in the climate are going to cause these incidences to happen more often.
So we cannot blame individual events, but we do know that these large term changes can be attributed or are more evident that things are happening. And obviously locusts, they eat a lot of these crops in Africa. And to make the connection with COVID, you know, coronavirus is causing a big change with regards to crops and the production of these different crops in America.
So it’s important to know that as the climate changes and as these big term big level changes happen, it’s going to result in these small level things that are going to start affecting our lives. That’s why it is just going to become increasingly important to know when those individual bad weather events are going to happen in order to prepare for these bad things and mitigate loss as we’ve discussed, but also we need to keep track of them.
We need to understand each year when these events are happening, how it’s trending over from these 10 years in the past and to these 10 years, how many bad hurricanes were there? How many snow storms were there? How much snow, how much rain, things like that.
So it’s just going to become increasingly important because what I think is going to happen is we’re going to get more of these weird weather things happening and then people are going to start to realize that the climate is changing.
Yeah, and we’re seeing that weather or just natural weather, everything is no longer normal. I mean, we can tie it back to COVID. Now in COVID-19 there’s been some statistics that up to 50% of USA crops are being led to rot on the field or being thrown out because you know, restaurants are closed and distributors don’t know what to do with their supply chain.
We have seen fortunately some good press and stories about, you know, different farmers in Idaho claiming that they are just going to open up free potatoes for all. And then people came from all these States to pick that up, but that’s not the whole story. Supply chains are being disrupted by Covid and it’s having a huge impact on crops. What’s your take, Jared, whether you guys are thinking about WeatherOptics on COVID and all that impact with weather and crops.
Sure. I think you make a really valid point comparing COVID to some sort of perfect storm that’s really taken the world by storm for lack of a better word. And we can imagine weather events themselves and by no means, should we be comparing the two because they’re completely different. However, they’re both instances where something we could call it relatively random occurs and has an impact on these things.
So I think the fact that coronavirus is being so disruptive across so many industries might even make people more interested in things that are analogous to it. In some ways, the weather can pop up relatively randomly and be quite disruptive across industries. As you mentioned, climate change is certainly something that we focus on and that we’re keeping tabs on. And we know that these crop yields are incredibly important to both everyday life in terms of consumption as well as stocks or the financial industry that trades things like the value of wheat.
And that is something where, when we started dipping our toes into the financial industry, people were excited about weather’s impacts on crops because they knew how important crops are to everyday life, as well as the economy. And one of the things that WeatherOptics has done in the past is acquire a company that had really good, similar analogous impact indices for these crops. And we’re implementing that into our portal because we know how important that is. And another next step is going to be implementing these crop indices, these crop impacts into those deliveries that I mentioned, because we know how important crops are.
And as you said, weather correlates with how well those crops are doing. So that’s a part of the business where we recognized how important that was. We took action to bring on individuals who had worked with that before. And it’s going to be a really large focus of the business. Moving forward is getting these crop indices up, testing their accuracy and deploying them across our product suite.
It’s excellent to really think about that because weather impacts all of us from the grocery store and the market on a daily basis and will continue for the foreseeable future. Speaking of the future, let’s talk about other trends. What else are you seeing that the future holds for us today?
Sure. Obviously in this day and age, everything is so uncertain and my hope for everyone that’s listening and for everyone in the U S and across the world is that we return to normalcy at some point. I’m a creature of habit, certainly and this has been quite disruptive.
And I know that people’s mental health is being affected by these circumstances, but I hope that people are able to grasp things that they care about and find work that’s meaningful and learn new skills and take this time that we have to further develop themselves because obviously the hope is that we come out of this learning something as a business and learn from these circumstances.
And we hope that this is a wake-up call not only just emergency preparedness as a whole and hopefully people will start rethinking how future prediction can be used, specifically within the machine learning world. Just considering that this is what we do, we’re predicting things that we don’t know.
I think this could even feed into that fire in terms of technology and improving and people realizing how important prediction is going to be. And then lastly, with regards to the technology space, a ton of people are obviously working from home. And I think for people like myself using github and using the cloud and building on the cloud, we’re going to see a big shift in our community.
In terms of how people are building and how many people are interested in this stuff, you know, you could have someone who’s working a job and then they go remote and that sparks an interest. For example, how does Zoom’s technology work? How do they have so many people making these calls at once? How is their infrastructure, how does this business run? And maybe it’ll just make people more excited about technology. I certainly hope so.
And taking this all home Jared, what call to action do you have for our listeners of HumAIn today?
I would say a good call to action would to be maybe once a week, check your forecasts and then just see for yourself maybe 4 or 5 days out, whether the forecast that you’re finding online is accurate based on what actually happened. I just think it’s pretty amazing that these days you can do a quick Google search and you can find at this day, at this hour, what’s the percent probability that it’s going to rain in my hometown.
And I think if people can use weather as a framework for technology and artificial intelligence as a whole, it will allow people to understand how powerful prediction is, how powerful the system that was built starting with Louis Richardson. And it might get them excited about the weather even if they don’t think about it or how it’s impacting them day to day.
Jared Goldberg, Head of Data Science, WeatherOptics, thanks for joining us on HumAIn.
David. This was really fun. Thanks for having me.
Thank you for listening to this episode of the HumAIn Podcast. What do you think? Did the show measure up to your thoughts on artificial intelligence, data science, future of work and developer education? Listeners, I want to hear from you so that I can offer you the most relevant trend setting and educational content on the market. You can reach me directly by email at email@example.com. Remember to share this episode with a friend, subscribe and leave a review on your preferred podcast app and tune into more episodes of HumAIn.