Oracle and The World Bee Project expand partnership following trials

How smart hives work

According to Simon Potts, professor of biodiversity at Reading University, understanding the health of honey bees is a lot like going to the GP, where a diagnosis will be made based on a range of measurements, like temperature and blood pressure.

“It’s a very complex and subtle set of factors that give a signature of whether a colony is healthy,” he says. “What we’re doing is using these sensors to collect a huge amount of data on all these different things, and then searching for the patterns across the ecosystem that will give us an indication of whether they are healthy or not.”

For example, using the acoustic monitoring sensor has revealed that an increase in the volume of low frequency noise means a colony is building up to a natural swarm.

With the forewarning this provides, beekeepers are then able to capture the swarm and start a colony in a new hive to increase both the amount of honey they can produce and crops they can pollinate.

Similarly, the analysis of acoustic data has also revealed a significant increase in noise levels is a sign of agitation in a colony. This means if a hive is attacked by a hornet, which can wipe out a colony by killing the queen and vast numbers of bees in just a few minutes, the system can send an alert to the beekeeper’s phone, who can then take appropriate action against the hornet.

According to Sabiha Rumani Malik, founder of The World Bee Project, the network’s goal is to generate previously unobtainable insights using data.

“Oracle cloud enables us to raise research to a new level,” she says. “We can’t have progress without data – we need evidence, we need knowledge.”

At the moment, each hive is producing roughly one million data points a day, meaning there are billions of rows of data for the whole network. This data is then streamed from the devices, where it is partly structured through edge analytics, into the Oracle cloud, where a further restructuring of the data takes place.

From here, a connector is used to transition the data into an autonomous database, which, according to Oracle’s project director and vice-president for cloud and innovation, John Abel, then allows bee experts to train an artificial neural network using machine learning (ML).

To deal with such a high volume and velocity of data, “the model has to be looking for the needle in the haystack,” he says.

However, as more and more data is fed through the ML system from across the ecosystem, the better it gets at identifying patterns, which can then be used for more accurate diagnoses.

The threat to bees

It is hoped this greater understanding will make it easier to protect the world’s honey bee population, which has been steadily falling since the 1950s. According to Potts, the honey bee population has been steadily falling since the 1950s, declining from roughly 330,000 hives to just over 100,000 at the turn of the century.

The threat has come from what Potts calls a “lethal cocktail” of factors, including the destruction of habitats, the increasing use of pesticides in modern farming, climate change, and the spread of pests and diseases.

“We have recently had a modest recovery, but we’re nowhere near where we want to be – or should be,” says Potts, who added that the economic and agricultural impact of losing bees would be disastrous.

“To put it into context, globally honey bees and other pollinators contribute pretty close to half a trillion pounds per year to food production, which is massive,” he says. “In the UK alone, that’s nearly a billion pounds a year.”

He added that, given 77% of all the food we eat depends on pollination by bees and other pollinators, there will also be a direct impact on human health, with vulnerable communities around the world being most affected.

The future of the project

Next on the agenda for Oracle and The World Bee Project is advancing solutions through in-depth scientific research, with 30 additional sites being added specifically for this purpose.

“We are injecting some pace into it by having more beacon sites in the UK, which will be named as we formally agree them,” says Abel. “We’re also setting up some specific location with Reading University to do proper scientific research.”

While he warned that the scientific research will take longer, as experts will want to interrogate a very high-quality data set before arriving at outcomes, he is hoping the data will be able to offer the experts unique insights into the life of bees.

“It’s an ongoing process,” he says. “We’ve got the proof of concept; now it’s really in the development and validation stage where we need to upscale it.”

Despite further fine-tuning and development, Potts says it is not just beekeepers and scientists that can benefit. “For growers, it can help them monitor and secure food production, and for governments, it can help monitor the health of the national pollinator population.

“We want to have this set of tools that all those groups can use, because they’re all struggling to get on top of honey bee health.”