Episode 5: How to make good use of data and digital tech

Vic Drought Hub - Farmland 1
Transcript

Kirsten Diprose:

Innovation Ag is made on the lands of the Gunditjmara and Wurundjeri peoples. We acknowledge the traditional owners of country throughout Australia. We pay our respects to elders past, present and emerging.

Dave Henry:

Everyone's talking about machine learning and AI, and I think it has a place and it has a role. And unless you understand the context within which it's being used and the use case, then there is a danger that it misinforms what you're actually trying to do.

Dwain Duxson:

There's lots of data in our system, but we haven't brought it to the surface yet and we're not using it to advantage. And by that, I mean to advantage our customers.

Kirsten Diprose:

Hello and welcome to Innovation Ag, brought to you by the Victoria Drought and Innovation Hub. I'm Kirsten Diprose. There are so many digital platforms out there promising to count your cows, measure your soil carbon and do your finances. It can be kind of overwhelming but also exciting. Personally, I quite like tech, so I love a digital solution, but when it comes to technology, just because you can doesn't always mean you should. In this episode, we are looking at how to assess what digital tech is useful for you, the concerns around who owns your data and how data and digital tech is changing the industry, from online marketplaces to real-time pest detection. But first, I'd like to introduce you to Dr. Dave Henry. He's a chief research scientist at the CSIRO in Melbourne. He currently leads the Climate Smart Agriculture and Digital Agriculture groups. He loves tech, but agriculture is his background.

Dave Henry:

I am definitely not a techie. People start going down the really technical route, particularly IT and I just glaze over. I'm very much an aggie at heart.

Kirsten Diprose:

David's main area of work these days is translating data, particularly climate variability data into useful and actionable insights. But he actually started, and I have to say, where a lot of good people in ag seem to start, and that's in agronomy.

Dave Henry:

I got out of Melbourne Uni way too long ago and I was doing a lot of work in pasture agronomy at the time and I started my science career moving over to Perth and working largely in the sheep industry over there. Also, a little bit in beef and dairy and export fodder. And in all of those, what interested me was there's a lot of technology, you know, things like satellite imagery. There was the start of that whole push towards putting sensors on farms and capturing data, but there was still a gap between how do we actually use the data, how do we actually get value out of that data to actually inform decisions? And it's sort of grown my interest from not just the livestock sectors but working for a while in the broadacre cropping area, and how we actually can take the learnings for using data and scale it across different sectors and how can we can learn from each other and how we can actually get value from that data.

Everyone's talking about machine learning and AI, and I think it has a place and it has a role, but unless you understand the context within which it's being used and the use case, then there is a danger that it misinforms what you're actually trying to do. So we really need to marry that technical side with the domain side.

Kirsten Diprose:

Dave actually did his PhD in pasture science, looking at whether technology could be used to better understand the quality of pastures.

Dave Henry:

So data can be visual estimation. I remember one of my early journeys was using satellite imagery to objectively measure pasture availability. And one of the reference methods for comparing that was actually people just walking around making visual estimations on how much pasture was available. And there's things in between that, of course. It's how do we make objective decisions and how do we use objective data to inform those decisions.

Kirsten Diprose:

I mean, we still do that now. We'll walk around a paddock and look at the pasture and... you know, because it's not just about how much grass you've got, obviously, it's about what kind of grass do you have? Do you have the magic clover and, you know, you get that little ruler out and... I mean, that's how I was taught to do it. Is that not how we do it anymore?

Dave Henry:

We actually still do it, but you can do it in other ways now. So people are talking about drones and aerial imagery. People are talking about satellite imagery and there are some products out there in the commercial world for estimating, for example, the amount of pasture. But what I would also like to make clear is I don't think these things replace the need for people to be interacting with their pasture, interacting with their livestock. Data should augment their decisions. A lot of the decisions that are made can be quite intuitive and based on experience and are very personal or driven by personal behaviours and business decisions and locations. So data has to augment or add to those decisions, and it might not actually change the decision, but maybe give them more confidence around that decision that they're going to make.

Kirsten Diprose:

Can you give me a bit of an example? So we're talking about pasture. Then when you look at sheep condition scoring, you can do that with your hand and do the tally up. Show me how you could use some sort of data entry or programme to increase profitability on your farm.

Dave Henry:

You touched on body condition score, and we can actually use imagery to estimate that objectively. So as animals walk through a race or over the scales, it automatically actually estimates the body condition. So why is that important? That might be important when you combine it with weight and age to actually say, "Should these animals go to market? What sort of price am I going to get at market now versus taking them back out to pasture and actually allowing them to increase their condition or change their weight?" So you're actually making a decision based on objective data as to, for example, are they market ready? Is this the right time? Am I going to get the right price? Versus, for example, having to buy and supplementary feed to get them to a different plane of condition. So that's an automated sort of system where data is potentially fed into a decision support tool. And obviously, there's more manual methods as well to put data into software.

Kirsten Diprose:

I think sometimes there can be a lot of software that's out there and it's hard to know which one to choose and they don't necessarily talk to each other and you don't want to spend all your time inputting data and especially the thought of coming back from a day's work and inputting data is not attractive to many people.

Dave Henry:

Yeah, this is one of the largest barriers I think in the ag sector, the requirement to manually input data into some applications and the, what we call, interoperability, so their ability to talk to each other.

You know, we've all got phones, we've got heaps of apps on our phones, how do we decide which is the best weather app for us? How do we decide which is the best sensor to deploy to measure soil moisture? I think storytelling is actually really important in this to listen to experiences, to be able to convey the value proposition of these bits of gear. But I think too, one of the largest challenges we have is ensuring that it's fit for purpose for that situation, for what that farmer wants to get out of it. And so we need to be better I think in actually providing evidence-based information, cut through some of the marketing hype to some degree, we're not just seeing in ag, we're seeing in all sorts of sectors, so that we cut to the chase about where is the evidence? If I purchase this technology, how do I actually get value out of it in my system knowing that my system might actually be different to nextdoor's farming system.

Kirsten Diprose:

We've heard in previous episodes, particularly the one on decision-making about really starting with the problem first rather than the exciting tech. So is it good to know what you need first and then shop for the thing that might solve it?

Dave Henry:

I challenged some of our cropping agronomists recently about soil moisture sensors in cropping systems and I said, "If you had an ability to measure soil moisture, what decisions would you actually do differently as a result?" So to try and get at that question, when do you need the data? How many sensors do you need? What sort of accuracy do you need? And it came down to some very distinct decisions that are made through the development of a crop where you've actually only got small windows of opportunity to actually do something different. So you've got to understand the use case, the decisions, and then look at the technology and how that fits to provide that data in the right way. I do think there is also room to think about the art of the possible. So we don't always know what technology is out there or what we might be able to do with the data, so there is room for that, but it really needs to start with the decision first.

Kirsten Diprose:

Most farmers really know their properties quite well and they know their paddocks well without that sort of data. They just know that this particular paddock doesn't seem to grow, you know, summer pasture well. So what can data give you that you mightn't already know?

Dave Henry:

Often a lot of the ways we collect data is indirect measures. A lot of our technology, whether it might be from aircraft or handheld sensors, might measure the greenness and it doesn't really necessarily tell you what the underlying factor is. And there's an opportunity I think to become more diagnostic about it. So instead of saying, "Oh, it might be acidic, or it might be saline, or..." Actually being more targeted at understanding what is the root cause, sorry for the pun, for that lower productivity in that corner.

So that's a diagnostic sort of example. I think the other one is understanding options for what you can do with that parcel of land. So let's just take the carbon story as an example again, or maybe biodiversity one is a really nice one. If I'm going to put certain amount of my farm back into native vegetation, where should I put it? Should I put it in that lower productive area, because the productivity impact's going to be less? You're still going to get the biodiversity outcomes. So actually understanding, yeah, sure, I've got a lot of history. I understand that northwest corner isn't as productive. But the question then becomes, "What if?". What can I do with it? What are the options that might be available to me? And data collected from a lot of different sources, not just on farm may help to actually provide that information.

Kirsten Diprose:

So that definitely ticks the box of a useful application of data. But in the future, this could go so much further, from monitoring what's happening on the farm now, to being able to model future scenarios.

Dave Henry:

I really think there's an opportunity to move beyond data for monitoring, sort of a situational awareness thing, "Oh, this is what's going on now," to actually being predictive and to be able to play out scenarios into the future. So this is where we can actually stack lots of data, but importantly, use either analytics or models to actually understand what future states might look like. Now that can play out in many different timeframes. We can think about, "Oh, how's the season looking?" So what's the situation like in three months' time? But it might also be in a climate change scenario that says, "What is the landscape likely to look like by 2050?" Now I know that's sort of, you know, 25 years out.

The finance sector's looking at this intently and using the forecasting, not just the data but the forecasting, to play out scenarios around what are my options. Now those options might be relatively smaller, capital investments on your farm, changes in land use, or they can be quite significant and that could be moving out of broadacre cropping into pastoral grazing systems, for example. So that scenario planning forecasting I think is a really important layer we need to put on top of data.

Your ability to forecast and your ability to understand how different decisions might play out into the future is a really important one to play out perhaps in a digital environment rather than jump in in the real world. And so we're seeing a bit of a push towards this idea about a digital twin. A digital twin is really just a representation of a farm or a region in the digital environment.

Kirsten Diprose:

Can you describe what a digital twin would look like? Does it go beyond the kind of 3D Google map of the farm?

Dave Henry:

In a way, it is, but just like on a real farm, you can drive around and you can interact and you can go and measure things and collect data. Imagine a 3D map of the farm that you can navigate around, but it's augmented with all those data points, not just from sensors maybe on your farm or soil carbon maps, but also climate information and market information. So you imagine this 3D environment. It's on your device, it might be on your phone, or on your computer. You can navigate around, you can interrogate data and you can run those 'what ifs' on top of it. So what if the climate projection was for a bumper season, how's that likely to play out versus just an average season? So running those scenarios. Bit of a gaming environment too, I think, Kirsten.

Kirsten Diprose:

Cool. It basically sounds like Minecraft, minus the zombies. But in all seriousness, the digital twin concept has much broader applications too. For instance, Ararat in Western Victoria is part of a digital twin project, with the council receiving state government funding to map the entire area with live sensor data from farmland to street scapes and roads. These tools can be used in flood management, for instance, or to better understand the impact of drought on an entire region. And of course, there are big companies interested in this space too.

Dave Henry:

Now think of the Googles and the Microsofts and the Amazon's digital twins, and we see that manifesting in things like Google Earth and Google Maps and so on. And so they're very much thinking about how they utilise their IT platforms as a foundation for people to build digital twins off. A bit of a technology push, perhaps. And it's important that we think about how we actually capture value from it rather than just, "Hey, here's another platform" and "Hey, here's some more data, and how you bring it together.

So interestingly, those big corporations, the IT ones are very interested is how they partner with companies to go to market with. So they want to provide the platform, the infrastructure, but the go-to-market strategy is actually through others that are more local and can actually utilise the platforms. From a more strategic broader sense, I think things like how do we retain vibrant communities in the face of climate events, whether they're extreme events, whether they're a change in climate, how do we actually just from say a government monitoring perspective, is how do we actually understand those regions which are going to be more at risk of climate events? How do we actually support communities? What's the likely impact going to be? Even things like changes in workforce. So if we're changing into a more technology-driven sector, does that mean that we need people with IT skills, with hardware and sensor skills?

Kirsten Diprose:

I know the federal government's partnering with groups like CSIRO and with the Weather Bureau to, you know, get some really good climate data and make that free. What's happening in that space?

Dave Henry:

Climate services for agriculture, which is government funded with CSIRO and Bureau of Meteorology, that is doing exactly as you described, taking what we know of climate data, and I say that loosely because it's looking back but also looking forward, and saying, how is that likely to impact and how has it impacted different sectors in different regions? It gets really complex because the climate impacts for a sheep producer may be quite different to the climate impacts for a wheat grower in exactly the same location. So climate services for ag is one example where they're trying to make it quite specific to different regions, make it specific to different sectors, so that the information is actually available specific to different land uses in different regions. So as an example of actually saying, "Oh, we know climate's important, but what do we actually need to provide that will enable people to make better decisions?"

Kirsten Diprose:

And that's really complex work when you look at the data and the tech, isn't it, to get that really region-specific information. It's going to take a bit of time to build it up.

Dave Henry:

It is, and there's some uncertainty, particularly we all look at our apps and think is it going to rain in the next 10 days? And so let alone thinking what's the seasonal forecast look like? So there's some uncertainty, but we can often package up the information in a way that is insightful, without being certain, but that can be hard to communicate and we're dealing in such a variable and volatile climate, it's hard to do sometimes. I think there's also an opportunity to integrate things like the broad scale climate data with on farm sensor data, because you can actually make it more localised during, if you like, a top-down, bottom-up marriage between these data sources.

Kirsten Diprose:

Well, that's the other area of variability, isn't it? Your soils can be so different from your neighbours'. So climate and the weather is only one factor.

Dave Henry:

Being able to bring together our best possible information on soils, climate and weather, land use management then becomes very complex. And I love this idea around how do you bring together farmer knowledge in that as well? So you're actually engaging early, you're listening to their problems, you're actually understanding what their pain points are and how best to tailor all those bits of data to actually make a difference.

Kirsten Diprose:

Yes, farmer knowledge. I love this idea of thinking about it almost like a dataset of its own. It's funny how it can be overlooked, though. One person who has been leveraging farmer knowledge for innovation for years now is Dwain Duxson. He founded Farm Tender, a national online marketplace made up of 65,000 buyers and sellers.

Dwain Duxson:

When I was growing up, I always wanted to be a farmer. With our chief stud, we had about 100 clients and I love the communication side, so I love communicating with all our customers. I'd send them weekly email newsletters and they really responded to that.

Kirsten Diprose:

While Dwain still enjoyed running the mixed farming operations with his family at Marnoo in Victoria's Wimmera district, he started to dream of something even bigger.

Dwain Duxson:

I wanted to reach more people and that's about the time when the internet was starting to come in vogue. So I sort of got a bit excited about that and I had a few plans to sort of exit my role at the farm.

Kirsten Diprose:

So where did the idea of Farm Tender come from? Did you launch straight into that or was there a period of time of playing around in other kind of arenas?

Dwain Duxson:

There was definitely a period of time. It was probably five or six years actually before we came up with the concept of Farm Tender and then it was a different looking business until that light bulb moment came along and it is like to what it is today. It's been the same sort of system for I think nearly 11 years now. But before that, yeah, we did fool around with a few different things, but they all sort of led to the culmination of Farm Tender, which is, you know, a online marketplace business.

Kirsten Diprose:

So tell me about those earlier days. Are we talking pre 2000? What era and what stage was the internet at?

Dwain Duxson:

I think I was 35. So I was born in '90, so early 2000s, it would've been. And with an agreement of the family, Paula and myself and the kids left the farm. So we drove out, we were living on the farm and managing the farm and then left, obviously left a bit of security. And then as we drove out the driveway, we went into sort of all the unknown and we moved to a place called Yarrawonga, which was 350 kilometres away. So you sort of had to go that distance to get away from the day-to-day running of the farm, because they just call you back to do stuff, you know how it is. So we went and lived there and started a couple of different businesses that had sort of online involved in them, but it wasn't until I sort of came across this online business that was sort of doing quotes and tenders and things like that, and I thought we could do that in agriculture.

So it started out as a buying service. So farmers would go and list what they require. So let's say they wanted a 200 tonne of MAP fertiliser, they would go and list that and suppliers would go and put a quote on, but it sort of never got going. The farmers loved it because they got different quotes off suppliers, but once a supplier lost four or five quotes in a row, they'd just say, "Well, bugger this. I'm never going to get a crack at this." And it sort of died that way. But we used to go to the field days and it wasn't until a farmer actually came up to me and said, "Why can't I sell stuff on your website instead of, you know, putting stuff I want to buy? And from that moment on, that was our pivot and it's just grown from there.

Kirsten Diprose:

I've only been involved in farming for the last 10 years, so I can't really imagine what it was like before then. So if you wanted to sell your header or something, you would list it, I imagine maybe in the newspaper. Is that what you'd do? Where would you advertise it?

Dwain Duxson:

Yeah, that was pretty much the only spot you could list it. And if it was a local newspaper, they only had a local coverage. If it was The Land newspaper, that was sort of only through New South Wales, or the Queensland Country Life or whatever. So you're sort of localised by where you could reach. But with the internet and with our business, we're getting eyeballs all over Australia on something. And what we find with farmers is they really know what they want and if they spot it, it doesn't matter where they are, they'll go and get it.

Kirsten Diprose:

Thanks to you, a lot of my holidays are dictated by what machinery we're picking up from around the country. So I don't mind. It gets the husband off the farm sometimes as we go to a trip to South Australia to pick up some sort of bit, some part.

Dwain Duxson:

It's amazing how far people travel to get the things that they really want and need.

Kirsten Diprose:

Dwain got in early when the internet was new, so I asked him, is it true that to be successful in AgTech, you have to be ahead of the curve?

Dwain Duxson:

I don't think it is. I suppose agriculture's a different beast again and can see why people coming from outside agriculture into agriculture find it difficult, because it's got a language of its own. And there's another big factor in agriculture that we are lucky enough to excel in is trust. So you've got to have a lot of trust with your customers, and you've got to build trust with farmers, and that can take a long period of time. And relationships and trust are a big part of our business. Our website and our app is our shopfront, and that's a digital side of it, but we have people behind the scenes helping facilitate the deals that is really the guts behind the business. And in agriculture, you can never sort of get away from that. That's always a constant.

Kirsten Diprose:

There's so much we can learn about digital tech by going back in time. Dwain mentioned how going to trade shows to talk to farmers is where he really understood what to actually do with the technology, which back then was just the internet, and he didn't know if farmers would travel the distances they do now for parts of machinery. That customer behaviour emerged alongside the tech. So his success was in one sense based on customer research, and another was probably a bit of luck by tech opening up an opportunity. But one thing technology can't do, well at least yet, is build relationships of trust, which is incredibly important in an online marketplace.

Dwain Duxson:

Our staff build up really strong relationships with our customers. So you've got a guy at let's say Wagga, who wants to come down to Warrnambool to buy something, you know, they don't know each other from a bar of soap. But in the middle our guys get involved and they say, "Well, we've dealt with Johnno down at Warrnambool for six years and what he says it is, it is." And that just gives that person buying another level of trust, you know, and the next level of trust because they sort of drop their guard and say, "Yeah, no worries, I'm happy to go ahead with the deal."

Kirsten Diprose:

What do you think the influence of that personal relationship that you say you need alongside tech? You know, I was in the city the other day and I hadn't been for a while and I realised I couldn't catch a train because I didn't have the myki. I wanted to go and order some food and then there was no number to call and I was like, "Oh." So I have to go on and use the app to order the food. Like, there's just no one to speak to these days. Can you get away with that in ag? Could you build a platform that works beautifully, that has great user experience, and there's no one on the other end, which would be a lot cheaper for you to build to not have to worry about paying a lot of staff for customer service?

Dwain Duxson:

Yeah, I honestly don't think you can and I don't think that's going to change anytime soon. So what we see is a lot of our older customers that have been with us since day dot, their sons and daughters are coming through, they're more tech-savvy, but they're still doing it the same way as their parents did where they rely on that communication and that phone call and relationship. So I don't think that's going away anytime soon. I look at livestock agents and probably 10 years ago, I thought they might've been superseded by now, but it's actually turned the other way. They're actually more important to a farmer's business now than they were 10 years ago because farms are very complex and farmers are very busy, so they rely on these people to do these jobs for them because they just can't do everything themselves.

Kirsten Diprose:

Let's pause for a moment to think about trust. That's some nice trustful music there. Dwain overcame his trust problems by supplying brokers to verify people and build relationships. And maybe you can create the feeling of trust with music, but people are having a hard time trusting big data right now. Recent research indicates significant levels of mistrust. One report found more than 60% of farmers surveyed had little to no trust in technology providers to maintain their privacy and not engage in unauthorised use of their data.

Dwain Duxson:

It's a big debate. Who owns the data and all this type of thing and whether it's secure. You know, you jump in your tractor and they know in headquarters what you're doing and what you've done for that while you're in that machine and working that machine. So they're recording that data and using it to make decisions to perhaps build better tractors. I'd like to think that that data that they're collecting and using is actually indirectly going back to you to build a better tractor or a more efficient tractor. So there's good sides of it and there's other companies that exploit it and just use your data and don't even acknowledge or ask or things like that. And the security side of things is a big issue, I think. And even down to people replicating invoices and things like that, there's things we need to address there to make sure that we don't get scammed. In our Farm Tender system, we do invoices on behalf of the seller to the buyer, and we sort of make sure that all those security checks are in place.

Kirsten Diprose:

In terms of data collection, what sort of data do you collect about your customers or people in the marketplace?

Dwain Duxson:

Just basic stuff. Like if one of our guys is on a phone call, you know, they're recording what they spoke to them about, and that's just more an internal thing for us to help. If that person leaves, you know, someone else can come in and pick up the slack and know what that person is about. I mean, we keep all the listing data, people that lists items and then we keep all the enquiry data for any people that enquire. And obviously, we keep contract data, we keep invoice data, but I suppose there's lots of data in our system because having been going for 11 years, we can go back to day dot of what sold and what it sold for and how many hours that tractor did, or what protein levels was that bit of hay, and things like that.

So there's lots of data in our system, but we haven't brought it to the surface yet and we're not using it to advantage, and by that I mean to advantage our customers. We know it's all there and I think that's something we can look at down the track is bringing that data to the surface. Yeah, there'll be lots more talk about it as we move, especially with AI coming in. AI relies on data, so it's got to be able to find data from somewhere to be successful.

Kirsten Diprose:

Yeah. What do you think the future uses of some of our data in agriculture will be with that added AI level? Because I think what you just said is a problem across the board, so you've sort of noticed you're collecting data but perhaps not using it.

Dwain Duxson:

You probably can't use it because it's dirty data, I suppose. You've got to be able to sort it before you can actually use it through anything that's useful. You know, they're talking about clean data and that's a well-used term, but it's sorting the data and it's the hard thing. Like as I said, we've got heaps of data in our system, but we haven't yet tried to sort it. Whether it's useful or not, I don't really know.

Kirsten Diprose:

But what we do know is that without adequate trust, the adoption of digital technology on farms is going to be adversely affected. So there's been quite a bit of work in this space. The National Farmers' Federation in consultation with the industry recently developed the Farm Data Code advocating for the fair and equitable use of farm data. Here's Dave Henry from the CSIRO again.

Dave Henry:

Ultimately the data that a farmer collects is theirs and they should be able to give the permission as to whether they share data or not. There are these basic privacy regulations around sharing a use of data.

So for example, when we work with farmers and collect data from farms, underlying principle: They own it and they have to give permission to share it for how it is used. But under the frame that data could provide a lot of opportunities if it is shared, and how do we actually create the environment to enable the safe and secure sharing of data, and if we need to anonymize it so that it can be used without necessarily pointing it back to individual owners.

Even as I said earlier, consumer pressures about understanding the environment in which livestock have been produced. Have they been produced in a sustainable way, in an animal friendly way? So that provision of data right through the supply chain might actually be really important for a farmer to communicate that their produce is being farmed in a responsible way. So that's quite a different use case. And we're seeing the finance sector, I think too, interacting with their clients in a way to inform lending decisions, for example. And that can be quite contentious about that sharing of data, particularly when it's around credit decisions.

Kirsten Diprose:

And data is being used for good to potentially solve some big issues. Meet Dr. Jessi Henneken, a research scientist at Agriculture Victoria, who is combining bees and data as part of the Smart Hive project, which is actually funded by the Vic Drought and Innovation Hub.

Dr. Jessi Henneken:

The Smart Hive project has two components. So the first is looking at whether or not we can track hives as they move from one location to another. So in order for bees to pollinate our produce, they get transported sometimes quite far distances and between different types of orchards. And we want to know if we can use trackers to not only check the location of the hive, but also to collect data on the health of the hive. So hive temperature is a really good indicator of hive health because bees are so good at maintaining the temperature of the hive.

Kirsten Diprose:

And the reason why the health of bees is so important is, yes, they're great pollinators, but turns out they're pest detectors too.

Dr. Jessi Henneken:

Bees are incredibly important to safeguarding our food supply. About one-third of Australian produce relies on bees for pollination, so that includes almonds, avocados, apples and pumpkins. So without bees, those industries are really going to suffer. And what's happening now is we've become a lot more global. So that increases the risks from the exposure of bees to these pests and diseases. And then we combine that with the threat that climate change poses. We really want to be safeguarding our bees now. We've actually started at our own smart farm. So Agriculture Victoria have an almond orchard that we use for doing all sorts of experimental work. So the hives that we are tracking and collecting samples from, started out at the Smart Farm, which is our own experimental orchard.

Kirsten Diprose:

What sort of digital technologies are used in this particular project?

Dr. Jessi Henneken:

So we're using trackers, so they're like physical devices that we've placed under the lid of the hives and they're collecting locational data and also data on the temperature and environment of the hive. So if you open it up, you would see your panels, chock-full of bees, hopefully. And then what we've done is we've attached two different types of trackers to the lids of the hives. So we might trial other types of trackers in the future, and as technology evolves, obviously, hopefully we can get ones maybe that are really specific for use in hives. And these trackers, they are physical devices, but they don't look all that technical, they don't have like wires and stuff coming out of them. They're just blocks of plastic that are a bit heavy that we've attached to the lid of the hive. But they're doing their thing, they're collecting data, they're sending them to our systems, and we can check on the hives and how they're going.

Kirsten Diprose:

The project is only in its early stage, but the hope is, it will allow growers and beekeepers to communicate with each other in real time, perhaps by an app which will be especially critical in the event of an outbreak of pests or disease. And it's all about early detection.

Dr. Jessi Henneken:

Assuming everything goes to plan, we want to have some sort of system where we can update the beekeepers on pests and diseases we've detect in our hives. And then they can take steps to either manage those hives. They might want to give those hives a break from pollination for a little while, or if it's something really exotic, we might then go into an emergency response mode where we make sure that that exotic pest isn't spreading into other hives.

Kirsten Diprose:

What's it like working with data and digital technology? You're a scientist, so I imagine you're not sitting there coding and doing all of the digital stuff. Do we need to be thinking more like this about, you know, merging fields and seeing how different industries and fields can really help each other?

Dr. Jessi Henneken:

Yeah, I think so. I think collaboration is always really exciting, and we're sort of at that stage where there's lots of opportunities to collaborate. The potential benefits of being able to do things remotely or from an app or the fact that you can just pull out your phone and maybe check on something that is happening at an orchard where your hives are or something like that. It's just going to give beekeepers, I guess, sort of real-time data on what's going on when they might not be able to physically be there. And that also helps everyone respond to things as they happen.

Kirsten Diprose:

This real-time data collection along with the benefits of artificial intelligence could be a game changer for the future of ag, but what would that future look like?

Here's that trust music again. Now it's imagination music. Join me, won't you in picturing a farm 20 years from now into the future. What do you see? Well, if you're like me, you saw a bunch of autonomous machines and drones whizzing around the farm while you sit by the pool, occasionally checking an app to make sure the farm is, you know, still running. Sounds like heaven, really, but how realistic is it?

Dave Henry:

You have to be able to trust that those AI and robots and the drones and all those sorts of things are doing the right things and making the right decisions.

There's a really strong trust that needs to be built for all of us and how we interact with these systems. And it's also important to note, I think that automation is only one outcome potentially of things like AI. AI might actually also be just a way of deriving insights from data, some of which, you might be able to automate, decisions. You might be able to make your tractor, your spraying decisions, your drone flights, and so on and so forth.

Automation, you know, as we've also seen, might be in the virtual fencing, is actually the automated movement of livestock around the landscape. So automation takes trust, and obviously we need to think very much around when the AI, for whatever reason, doesn't quite follow the path that we want it to follow. I think one of our challenges when we think about AI and machine learning is if we have the data, it can learn, but it has to learn from data to be able to forecast. And so it might actually be constrained by the availability of data in enough scenarios to make reliable predictions moving forward. It's still an evolving field, and we may in fact be constrained to some degree about the amount of available data for these wonderful AI things to actually learn from.

Kirsten Diprose:

For Dave, the most exciting role of data, at least in the near future, lies in the ability to bring it all together.

Dave Henry:

There's a lot of data collected on farm by individuals, by companies, all through the supply chain, by other means such as satellites. What excites me is with advances in IT platforms, with advances in analytics, whether we call it machine learning or AI or deep learning and all these sorts of things, our ability to bring it all together and actually derive insights from a predictive sense, not a retrospective sense, to derive insights from all this federated data that we've never been able to do before. So it's bringing data, it's bringing in models, is bringing in analytics and decisions all together in one place, and actually taking advantage of all this data, which is loosely at the moment, quite disconnected.

Kirsten Diprose:

And that's it for this episode of Innovation Ag. Thank you to our guests, Dr. Dave Henry, Researcher Leader in Digital Agriculture at the CSIRO, Dwain Duxson, the founder and CEO of Farm Tender, and Dr. Jessi Henneken, Research Scientist at Agriculture Victoria. Our next episode, number six, looks at how to stay operational while making major changes on farm or in AgTech. You can find the episode transcript on our website, vicdroughthub.org.au.

Thank you for listening. This episode is written and hosted by me, Kirsten Diprose, produced by Rachael Thompson. And we have editorial input from scientists, academics, and farming groups involved in the Victoria Drought and Innovation Hub. This podcast is funded by the Australian Government's Future Drought Fund. Catch you next time.

 

Innovation Ag Ep05 tile

There are so many digital platforms out there, promising to count your cows, measure your soil carbon and do your finances. So how do know which data collecting applications are going to be most useful?

From GPS-driven tractors, online auctions, weather and climate tracking apps to soil sensors, there is an increasing amount of data available to farmers and agribusiness. But are we actually using this data to the best effect? Data can unlock new opportunities, from greater market access to better knowledge about crops and pasture that can lead to higher productivity and better environmental outcomes.

However, research also shows there’s significant mistrust amongst farmers about the collection and sharing of data. And it’s these concerns which inhibit technology adoption. A lack of data integration can also turn people away (no one wants to use several digital applications, that don’t ‘speak to each other’) So how do we ensure our data is being managed correctly and effectively? Also, what data should we be collecting NOW to be ready to access future opportunities?

Guests:
Dave Henry, Chief Research Scientist/ Research Leader, Digital Agriculture, CSIRO
A Chief Research Scientist at Australia’s national science agency CSIRO, Dr David Henry currently leads the Climate Smart Agriculture Group within CSIRO’s Agriculture & Food Business Unit. Spread across five Australian states, the team has an innovative focus on Climate Adaptation and Sustainability Assessment and Metrics.

Dwain Duxson,  Founder of Farm Tender
Dwain started Farm Tender 11 years ago and it has evolved into an online Marketplace for Ag with over 65,000 member and throughput of $100 million plus in total sales.

He started Farm Tender while he was still running mixed farming operations with his family at Marnoo in Victoria’s Wimmera district.

Jessi Henneken, Research Scientist, Agriculture Victoria Research
Jessi is a research scientist with Agriculture Victoria Research (AVR).

She is currently working on two projects: a detection study on grapevine phylloxera and the Smart Hives project.  She is excited about AVR’s research in integrated pest management, which aims to develop more sustainable and environmentally friendly crop protection practices.