Blog

Podcast: How geospatial data analysis can track pandemic impacts on food supply chains

Bryan Yates, general manager and director of sales for Europe, Middle East and Africa at Orbital Insight, was interviewed by Ian Welsh of Innovation Forum about how analysis of geospatial data can help companies track their supply chains and gauge risks. The transcript of their conversation follows, but you can also listen to the episode here.

Ian Welsh: Hello and welcome to the Innovation Forum podcast for Friday the 17th of April with me Ian Welsh. I hope you and yours are well. Earlier this week I spoke with Bryan Yates from Orbital Insight, who gave some fascinating reflection into how analysis of geospatial data can help track the spread of the coronavirus pandemic and measure effectiveness of, for example, isolation measures.

It can also give companies important detail about supply chain robustness and potential new pinch points as suppliers cope with the new challenges that pandemic has created. That's coming up but first some food business news:

Food supply security and animal welfare look set to be big issues over the coming months. Not at least because of the controversy over oversew cold wet markets in Asia where alive wild animals are bought and sold, and often skulled conditions, one of which of course, has been linked to starting the COVID-19 pandemic through viruses spreading between species and to humans.

The latest business benchmark on farm animal welfare covering 150 of the world's largest food companies, highlights concerns about the wellbeing of farm animals in Asian food supply chains. Of the 17 Asian companies in the benchmark 15 are on the bottom tier because of a lack of any detail on how they manage risks associated with farm animal welfare, pig poultry and seafood production, and of course a big business for Asian companies.

And two areas of particular concern, experts say are caged hens for egg production and pregnant pigs in crates. The benchmark’s authors argue that farm animal welfare is an increasingly important driver of business value and investment risk, and companies should therefore publish commitments to improve farm animal welfare through business operations and see how they will improve them and report on performance.

Child labor and the cooker sector is coming under scrutiny, with a new U.S. government report apparently set to highlight an increase in use of children on cocoa farms in the key producing countries of Ghana and Cote d'Ivoire. Roger says that a draft of the report contains details of more than 2 million children working in the sector in the 2019 season, which is higher than in 2010 when companies, including Mars, Hershey, Nestle, and Cargill pledged to reduce the worst forms of child labor, 70% by 2020.

While there's certainly been some hard work to tackle the extremely challenging problem given the nature of the small holder farmers involved the report will undoubtedly be disappointing for cocoa Watchers. They were all Cocoa Federation, which represents the sector has acknowledged that the industry is not yet on track to meet its 2020 targets.

An impact of the pandemic has been perhaps inevitable uptick in the use of plastic alongside a slump in newcycling capacity in the midst of the crisis, a number of plastic bag bans had been reversed, and with oil prices dramatically law, a perfect storm for a single use plastic is developing, at least for now.

One positive though is the news of the development of the use of a bacterial enzyme to successfully break down PET plastic bottles. Research published in nature magazine reveals that the enzyme originally found in composting leaves can break down PET bottles in a matter of hours, leaving material that can then be used to make new high quality food grade bottles.

Part of the process will for now will still cost a bit more than creating virgin plastic. Lower quality recycled plastic does already sell at premium prices because of the lack of supply. And while this new process only tackles PET and not other types of single use plastic, it is certainly typical of the sort of innovation that is required as part of the mix of solutions to properly get to grips with the plastic pollution problem.

The next events in Innovation Forum’s spring conference season are rapidly approaching. We've been working to create interactive, engaging online conference experiences with not only the usual engaging content from plenary sessions and breakouts, but also some exciting opportunities for the networking and conversations on the site of the main event that are so much of the value from conferences.

We'll deliver increased networking time as well as dedicated slots to host group or one-to-one meetings and discussion with relevant stakeholders. We know that there are constraints involved in online working and we'll reflect these in how the events are presented. So on the 27th, 28th, and 29th of April, we'll be meeting to discuss sustainable apparel and textiles.

Among the experts joining us will be representative of Primark, H&M, Tommy Hilfiger, CNA, and Eileen Fisher. Special price tickets are now available and I do hope you can join us. Please also joined us online on the 5th and 6th of May for the ethical trading human rights forum being held at U.S. East coast times.

If you attend, you'll hear from senior representatives at Walmart, Microsoft, Intel, PepsiCo, Hilton, Tyson Foods, Timberland, and many others. We're also developing how we will present our Future of Food event series on the 27th and 28th of May and the 2nd and 3rd of June. We will be in touch with all delegates directly, of course, but the best place to keep up with the details is to go to the relevant event pages at innovationforum.co.uk.

Coming up now are highlights from our conversation I had earlier this week with Bryan Yates, General Manager and Director of Sales for Europe, Middle East, and Africa for Orbital Insight. Bryan and his colleagues have been working on developing the use of geospatial data to track the impact of the COVID-19 pandemic and supply chains and resulting behavioral changes.

It's fascinating stuff as I hope you'll agree.

Why don't you start off, Bryan, by giving us a bit of an introduction as to what Orbital Insight does.

Bryan Yates: Orbital Insight is a Silicon Valley startup. We've been around for a few years now, and really we're what we term as a geospatial analytics company. We take a whole host of geospatial data sources like satellite imagery, terrestrial data from cell phones, AIS transponder information from ships, different types of flying objects like drones and things like that.

And what we do is we help our clients use those very difficult data sources to answer questions about the planet. So our mission statement is helping our clients understand what is happening to and on the Earth. And we've brought to bear using our technology, these very unique data sources that are not generally available to the commercial markets or to a lot of NGOs and public sector clients and make them available to ask these questions of the planet.

Really our goal is to help our clients make better decisions and have a better understanding into opaque areas of the world, and to help track activity as well as the development of the planet, and then the activity on the planet as well.

Ian Welsh: One of the reasons we're talking right now is that I know you've been tracking impacts of the Coronavirus epidemic, particularly around transportation and movement of food supply.

What have you been doing specifically to track these things and how has this been useful for your clients?

Bryan Yates: One of the areas we've been very focused on during this unfortunate situation, this pandemic, is helping our clients understand their supply chains a little bit better, or helping our clients understand supply chains of the companies that they either invest in, or they're more intimately involved in or interact with or do business with. Well, that's been a big focus for us. So everything from food supply to the supply of oil and energy related products around the world. We've been helping our clients understand how their supply chains may be impacted, how the supply chains of their key clients may be impacted so that they can make very quick operational decisions during this time, to help with the end goal of getting their products, be it food, be it respirators or healthcare equipment out to the market to help alleviate out of the crisis or to get people fed during this time where there's massive disruption.

So that's a really big piece of what we've been doing. We've also been helping our clients understand what consumer activity is going on. Are people still spending? Who are the retailers or areas of the world that are most hard hit by this pandemic, and what's a consumer activity that has really sort of waned off? We're also helping our clients more at the public sector and NGO level understand things like the impact of social distancing, which particular cities are practicing at well, which are not.

I'm looking at big transportation hubs and the effect on those. And then as it relates to kind of the global energy markets, we're also helping our clients understand the impact here in terms of supply and demand for oil and other related products. So lots of different things that we're able to use, our capabilities to help our clients understand during this time.

Ian Welsh: Let's pick up on a couple of those. And you're seeing that you're monitoring cities that are social and those that aren't. What typifies a city that is doing it well and what typifies a city that’s doing it badly?

Bryan Yates: I don't have examples off the top of my head. I know San Francisco is something we've looked at just given that we're right there in the Valley and they've been practicing the social distancing very well, but it's also been related, unfortunately, as we've seen probably a lot on the news, all of us to the type of job and the demographics of a particular area of a particular city.

So for instance, in New York city, we're seeing hubs that service the more wealthy areas like Westchester and New Jersey. So places like Penn Station and Grand Central, having massive dips in consumer activity and folks coming through those transport hubs. But places like the Staten Island Ferry, which services generally Staten Island and surrounds being still quite busy.

So there's definitely been an impact to areas that are not able to work from home and do things remotely, and we're still seeing activity in places like that. So quite an interesting finding that we saw there.

Ian Welsh: In terms of food supply, are there any particular pinch points that you've identified or food supply chains?

Byran Yates: Well, I think a big topic for our clients and places like the Unilevers of the world has been a better understanding of port traffic. So we leverage AIS data, which is a universally accepted method of transponding information from ships for safety reasons so that they aren't collisions at sea. But that data is available to companies like ourselves to help understand the length of time it's now taking for ships to get into port and for things to get unloaded.

And we're seeing major pinch points at ports around the world where there's huge delays in terms of ships having to anchor for many days outside of what their usual norms are before they can actually come into port and get unloaded. And then we're seeing obviously also delays at those ports just in terms of lack of workforce there.

So we're able to detect using our cell phone and geolocation capabilities to understand anonymized worker activity. So we're seeing pinch points at the ports as well as once it gets to port and goes out to the distribution centers, there's a delay in terms of getting those goods out into the market and out into where they need to be.

That's one of definitely the areas we've seen where there are certainly some disruptions and certainly delays.

Ian Welsh: And I guess as we go through the pandemic, go through lockdowns in many parts of the world and come back to the other side, monitoring of that data to become ever more important to see where resources can be put in place, perhaps to help ease a pinch points.

Bryan Yates: Yeah, of course. I mean, there's efforts by a lot of our clients to start to A) Understand their supply chains a lot better. So when things are humming along the supply chain intelligence piece of what we do is certainly crucial because they want to plan and make sure operations are running smoothly. But when you're dealing with difficult times like this, it becomes even more critical to not only understand your supply chain directly, but to understand deep into your supply chain. So who are my suppliers, suppliers, and how may they be affected as we come out of this in order to maintain normal operations Clients are having to understand A) what they've classically used as suppliers and if those have been impacted or they shut down or there disruptions there?

Then having to understand and make very quick operational decisions to keep that supply chain going. So getting quick information from tools like ours into their procurement teams and into their supply chain planning teams so that they can pivot and move if they are seeing big disruptions at critical suppliers so that they can go to their plan B and plan C and then insure normal operations. It's giving them that intel that they never had before and now is critical to them making decisions in order to keep their supply chains resilient.

Ian Welsh: Let's think a little bit more about behavior change. I know there's something you've been monitoring. I mean, there was surprising behavior changes that you've seen over the course of the pandemic?

Bryan Yates: Things that come to mind are airports. So we're seeing in the U.S. we do a lot of tracking and we're seeing obviously air travel fall off the cliff and you're seeing all airports being impacted. But as the pandemic spreads across the U S we're seeing different timeframes. The impact of those airports are in Los Angeles, for instance.

We're seeing a massive drop off in traffic, but in places like Orlando, we haven't seen as severe a drop off. I assume that's just because Florida was later on in terms of getting a lot of these pandemics, or we're seeing a lot of people fleeing to places where they either have second homes or where there's areas that are less impacted than where they are.

So that's one example. Other things that we're seeing, gas stations and things of that nature. Right? So we were able to track with our technology, hundreds of thousands of areas of interest or places like gas stations around the world. And we're seeing again in the U.S. an example where seeing real decline in the activity at the gas pump for areas that have put the shelter in place in time or early on versus others that aren't. Just interesting things in with regard to pattern of life and to consumer behavior.

Ian Welsh: The impact of the lack of air traffic is extraordinary. I mean, I live in the West Oxfordshire-Oxfordshire border and you can see the entire sky from horizon to horizon, but a single country on it. That's something I've never seen before and I'll never see again in my life in a day. So it's extraordinary decline in air traffic,

Bryan Yates: We’re also seeing things like in South Korea and around the world. We're able to detect and understand activity on the road using our algorithms to track cars and trucks, and we're seeing massive declines, as you would may expect in cars on the road, but still many trucks or even an increase in the amount of trucks on the road just to deliver critical services and things of that nature.

It’s been very interesting using those capabilities to see not only how the pandemics affected, but countries like South Korea and China that are starting to come out of this, who's coming onto the roads first, what has been the rebound in terms of traffic and things like that. And we're certainly seeing in places like Beijing traffic increasing by magnitudes of hundreds of percents on the road. You know, very quickly off to some of the lockdown measures have been removed and eased.

Ian Welsh: Let's think a bit more about how your technology works more generally. You work with companies to monitor their supply chains to increase sustainability of their supply chains. Your company says that your approach specifically delivers the benefits of traceability, sustainability, and brand equity. So can you talk a little bit more about how your technology does that and how it will be working as your clients come out of the pandemic in the coming months?

Bryan Yates: A lot of the supply chain work we've done historically has focused on areas like sustainable sourcing and traceability, which is for commodities like Palm oil and soy and such.

Obviously, he'd been a very important piece of all the sustainable initiatives that brands like the Nestlés and Unilevers of the world have committed to. So we've used those capabilities to help them map and understand deep into the supply chain. Because I'm sure, as you know, as part of the Innovation Forum, that especially in opaque markets like Indonesia or Brazil, there's a huge lack of visibility as it relates to supply chains deep down into tier two three and then all the way down to the small holder or the or the farm level, which has been a huge pain point for the industry in terms of their traceability initiatives and their traceability efforts and their commitments to traceability.

Our capabilities have helped them, and it's been cool that they've been very extensible to tracking supply chains, not only upstream or downstream as well. But to your question, what we do there is we use our terrestrial data from cell phones and mobile phones. That again, is very much anonymized and GDPR and CCPA compliant, but we use that data to track and understand patterns of life. So we're able to take, for instance, a mill or a crusher or a collection point in Indonesia or in the Amazon or in other parts of the world, and we're able to say, okay, based on this particular area of interest tell me who's touched that. And we've got a lot of data science that helps us understand who a truck driver may be or a trader coming into these areas. So tell us who’s in this facility and what has touched this particular facility. So we call that supply chain discovery or mapping. So helping them understand not only, okay, I know that I'm buying from Cargill and Bunge, but who's coming into those facilities, which has been previously a difficult area for clients, and they don't really have much visibility into that.

And then the second thing is, once we start to help them map out these facilities and this supply chain with more clarity, we're able then to track these devices as they move between these particular facilities. So we're able to quantify how much of an impact that say, hey, somebody's going to a protected area in the Amazon bringing in goods to a particular crusher. How many trips are we seeing there? And that becomes then a proxy for how exposed they are to those particular suppliers. Helping them really trace and understand that and then react accordingly. So if they're seeing activity coming from areas that are other protected forest, then they know that they, hey, they need to go in and investigate that a little bit more.

And what we underly all of this with is our capabilities on deforestation. So we've been working with WRI for many years to help them understand where planted versus natural forests come up around the world. So we use our computer vision and AI algorithms to detect planted versus natural forest. So as you can imagine, that becomes a foundational base layer to this investigation that I just explained previously.

So you use that as kind of a map to understand where your problem areas may be. And then we use our capabilities on discovery and quantification of supply chains to understand the movement around those areas and if there are any risky areas around. That's our capabilities in that front.

And as it relates to brand equity, I think we all know just consumer behavior is changing. People are much more conscious of where their goods are coming from. And sustainability has become not a nice to have, but a necessity for companies like Unilever these days. So they've been very much all in on this and working with us very closely to help co-develop these capabilities so that they can live up to their promises and their mandates for the market and for the globe that consumes their goods.

Ian Welsh: This technology is all absolutely fascinating. So can you tell us just what is the process of identifying planted versus national forest look like?

Bryan Yates: What happens is we train computer vision algorithms based on huge amounts of data inputs. So a computer vision algorithm is trained similar to how a child learns.

So the more and more data that we feed into these algorithms the smarter it becomes in terms of identifying these things. So the end goal and what we do is these algorithms get trained to look at satellite images, like a human would look at them and are able to then identify, okay, a planted forest is much more in a straight line or in certain shapes as opposed to what a natural forest would look like.

Over hundreds or even thousands of images that train these algorithms, the algorithm then is able to detect this, and the idea is really to do it at scale. Classically humans, and the way that governments and institutions have used satellite data in the past is to actually have analysts just look at the images and understand what's going on and make determinations.

But there's so much imagery coming down and so much more imagery available. And as you can imagine, the swathes of land are just massive. And being able to detect and understand all of this at scale and in a timely manner is just not feasible for a human being to do. So you need computers to help you do that.

So we train these algorithms to help identify these particular areas of planted versus natural forest. And then as I said, that becomes a very important base layer to start to understand when you track movement around these, where potential problem areas are. What's really cool about our capability is that there's products out there, and again, we work with WRI and their Flobal Forest Watch project and things of that nature, but you're not really able as a Unilever or a P&G to understand, okay, how is that directly with that fire or that particular deforestation alert that I received, how is that impacting my own supply chain?

So our capabilities on top of that, help them understand: Oh wow, okay, I bought from that particular Cargill mill, and there's been a lot of activity going on based on coming into that mill from this area that perhaps shouldn't be going on. We're not here to be the judge or the jury, and we're just here to provide information so that these companies can then collaborate and work to fix these issues.

Obviously, as you know, it's a very complicated scenario and there’s varying reasons why folks are acting in the way they're acting in these, especially in places like Indonesia and Brazil, and we're not there to judge. We're just there to provide more information and more clarity as to what's going on so that these companies can go through their mitigation methods and know essentially where to look and where to concentrate their resources.

Ian Welsh: This Is obviously a fascinating area and is fast moving and that the changes in the use of geospatial data in an extraordinary well, what's gonna be next? What do you think is coming down the pipe? Where will this go in the future?

Bryan Yates: We have a platform called go that enables all of this investigation. Where it will go in the future? I think we're just going to get more and more sources, especially now. I think it's become quite apparent that we're not going to be able to move as freely around the world and we're going to need things from space or high altitude drones or more automated procedures to monitor supply chains, to monitor activity of people, to understand the Earth a little bit better without traveling to those places. So there's always been an efficiency play, right? Because it's difficult to get to certain areas of the world and you can't get there that much. But I think that's been exacerbated, given the situation and the crisis that we're all living through, there'll be more and more ways that we'll be able to utilize our technology to help clients monitor their assets, their supply chains, and the consumers that they need to interact with remotely.

So it's just developing those capabilities, getting our platform more and more developed so that the data gets to people quicker and quicker. And then bringing in new data sources as always, is top of mind for us and is always something that we're doing. So as more and more of these data sources come available, we'll start to ingest those and use those for our clients as investigation and for them to be able to help understand what's happening to and on the Earth.

And a few examples of those are connected car data. So we just ingested millions of connected cars around the United States to help understand the movements there. There's high altitude drones that we're working with our partner Airbus that you can put up into semi orbit and have sit there for months and months and monitor areas at scale.

So it's really more and more data sources and just keeping on building out our capabilities and technology and algorithms to help our clients understand what's going on on the Earth better related to the assets that they care about.

Ian Welsh: Well as you say, who knows how things are going to be once the effects of the COVID-19 pandemic clear out? But it certainly would be very interesting to see how they use of geospatial data develops in the future, but for now, Bryan Yates from Orbital Insight.

Bryan Yates: Thank you very much. Thank you and appreciate the opportunity.

Ian Welsh: As ever, there's plenty more insight and analysis available online at innovationforum.co.uk including the usual podcasts and interviews.

And if you're attending either the upcoming sustainable apparel and textiles or ethical trade and human rights events, I encourage you to familiarize yourselves with the virtual conference platforms that we've been working hard to develop. And I look forward to meeting you online in the coming weeks, but that's all for now.

I’m Ian Welsh, and I'll be back next week.