I spent a very interesting, and pleasant day in Tasmania last Friday courtesy of Microsoft and IoT startup, The Yield, learning all about how The Yield is using IoT technology to help Tasmania’s oyster farmers be more productive, more efficient, and hence more competitive.
Along the way I also learnt something about what’s involved in oyster farming and believe me, it ain’t easy. Aside from the various hazards that The Yield’s IoT systems are designed to alleviate — of which more anon — producing those tasty uniform oysters elegantly served in up market restaurants is no mean feat. It requires constant nurturing, grading and shifting of oysters in their beds. Otherwise it’s survival of the fittest: you end up with a few monster oysters that suck all the nutrients out of the seawater and lots of shrivelled and undernourished specimens that no self-respecting chef would dream of serving up.
That’s just one aspect of it. Suffice to say I will never look at a plate of oysters in the same way again, but back to The Yield and IoT.
The Yield’s service to Tasmanian oyster growers addresses two problems. When heavy rain washes water from the land into the estuary, pollutant levels rise and oysters absorb contaminants making them dangerous if consumed. Government regulations prohibit harvesting of oysters until it’s judged that tides and river flow have washed away the contaminants and the oysters have cleaned themselves.
Historically the decision to cease or resume harvesting has been made by The Tasmanian Department of Primary Industries, Parks, Water and Environment based solely on rainfall as measured at weather stations the might be many kilometres from the estuary, and naturally has erred on the conservative side.
The Yield’s technology is able to predict when closure is necessary much more accurately. According to The Yield, weather closures have cost the Tasmanian oyster industry an average of $4.3m a year for the past three years. On this basis, if its IoT technology delivers just a 30 percent reduction in closures it would return four times the investment.
The other big problem facing the Tasmanian oyster industry is Pacific Oyster Mortality Syndrome (POMS), a viral disease that can wipe out almost entire populations. It’s not fully understood but what is known is that outbreaks can be triggered by higher water temperatures, so if oysters can be removed from the water ahead of a rise in water temperature they can be saved.
What I learnt is that IoT is really only one component of what The Yield offers. Its real value proposition, and where its key intellectual property resides, is in predictive analytics. As CEO and founder, Ros Harvey puts it: “We are in the business of managing uncertainty.”
The only ‘things’ in The Yield’s system are solar-powered salinity sensors that it has deployed at a number of locations in the Derwent River estuary, designed by its shareholder, Bosch and incorporating a highly accurate and reliable salinity measuring device. It combines this salinity data with publicly available, temperature, barometric pressure and tidal data.
Harvey says there are four components to The Yield’s system: Sensors, platform, data analytics, user interface. It’s using Microsoft’s Azure cloud services and the Azure IoT Suite to host its platform, but Harvey says, “The value proposition is in the data analytics.”
She explains: “We take the tidal signature and offset that for harmonics, for the shape of the bay and for weather and we get this highly accurate tidal signature or prediction curve. And because salt and fresh water don’t mix, when you measure salinity at a single point you get a salinity signature that echoes the tidal signature.
Patent prediction algorithms
“So where the smarts are is in the analytics. When we get a rain event in the catchment we can predict the local salinity as rainfall flows through the catchment. This is what we have developed the algorithms for and that are subject to patent applications. We can do a pretty accurate prediction.
“We developed the mathematical models with our data science team and to scale those we employ machine learning so the model is rapidly scalable when we want to apply it to different bays.”
Temperature prediction is particularly important to alleviate the impact of POMS, Harvey says. “What the industry is doing is window farming: they are growing oysters in areas that are not infected by the virus they have to get them out of the water before the temperature rises and turns the virus on again.”
While the oyster application might seem impressive it is, according to Harvey, not the main game. “We see oysters as our entrée,” she says. “It’s where we are looking at how to demonstrate the business model and the technology and how we go to market, but really it is platform technology that we are demonstrating here.”
The Yield’s story is another example of how the real value in IoT is not in networks of things feeding in data: that aspect is likely to be rapidly commoditised. It’s going to be in gathering disparate sources of data, discovering relationships between them, providing new insights and, best of all: predicting the future.