SAS creates tools to help count world’s bee population and monitor for murder hornets
Times have been tough for the honey bee in recent years.
Beekeepers across the United States lost 40.7% of their colonies from April 2018 to April 2019, a higher-than-average loss rate, according to the nonprofit Bee Informed Partnership, as things like habitat loss, climate change and pesticides killed the fragile creatures responsible for pollinating billions of dollars worth of food plants.
And if that weren’t enough, so-called “murder hornets” have now entered the United States.
Now data analytics giant SAS is hoping that its technology might prove useful in helping monitor the future health of the world’s honey bee population.
In partnership with a group at Appalachian State University, SAS has created a visualization tool to count the world’s bee population, creating a map and database of hives.
The tool is meant to allow individuals from around the world to tag hive locations and upload photos of them. With enough entries, the groups behind the effort hope that a large enough data set will emerge that universities and research institutes can mine it for insights on how to better protect the creatures.
“SAS’ data visualization will show the crowd-sourced location of bees and other pollinators,” Appalachian State professor Joseph Cazier said in a statement. “In a later phase of the project, researchers can overlay key data points like crop yield, precipitation and other contributing factors of bee health, gathering a more comprehensive understanding of our world’s pollinators.”
SAS has also developed machine learning and visual analytics tools to gather real-time data from bee hives.
Honey bees have become a passion project for SAS, with the company hosting four hives on its campus. The hives are taken care of by Bee Downtown, a startup that maintains hundreds of beehives on corporate campuses around the Triangle, creating a connected ecosystem of colonies that stretches 60 miles from Garner to Chapel Hill.
SAS researchers have developed a bioacoustic monitoring system to track real-time conditions from beehives without disturbing the insects. The monitoring system — tested on SAS’ four hives — can track hundreds of data points that could give a beekeeper more time to save a teetering population or help a healthy hive grow.
“We can stream data from hives in real time and analyze those for red flag incidents that beekeepers would need to respond to,” Sarah Myers, a product marketing manager at SAS, said in an interview.
The program can monitor a hive’s weight changes, its temperature and humidity levels, flight activity and acoustics.
“We can capture audio, and if we hear the queen is distressed that might mean she needs to be replaced,” Myers said.
Through machine learning algorithms in the company’s Viya software, SAS can analyze images to determine what gender of bees are coming out of the hive, how much pollen they are bringing back in or whether an invasive species — like the the Asian giant hornet, a.k.a. the murder hornet — has entered a hive.
The Asian giant hornet was first sighted in the United States in Washington in December, according to the Washington State Department of Agriculture, and earlier this year, an article about the insect in The New York Times went viral, as many saw a mixture of humor and despair in the fact that murder hornets and COVID-19 both surfaced in the U.S. around the same time.
Despite worries that the hornet could spread throughout the U.S., entomologists say the chances of it showing up on the East Coast are slim, The Charlotte Observer reported. The hornet is often confused with other insects, like cicada killers and the European hornet.
“As long as you have a large image bank of the Asian [giant] hornet versus the honey bee,” Myers said, “you could train the model to know whether it is a honey bee or a murder hornet” that just entered the hive.
SAS has been testing the monitoring system on its campus since last April. The monitoring system uses the same machine learning program that SAS has used in other applications, like assessing Wake County property values in real time or helping police departments monitor traffic incidents and crime.
“We definitely now have a good understanding of seasonal activity,” Myers said. “When flowers are blooming, the hive is growing in population and honey is increasing. But how quickly that happens you wouldn’t know unless you have data to see.”
“In one of our hives,” she added, “you can easily see that in a week’s time when nectar is flowing that hive can gain 20 pounds in a one-week period. As a beekeeper [myself], when I see that I think I might need to add more equipment or take frames of honey away” so they can produce more.
SAS said its research team is continuing to look for ways to broaden the use of technology to help bees.
This story was produced with financial support from a coalition of partners led by Innovate Raleigh as part of an independent journalism fellowship program. The N&O maintains full editorial control of the work. Learn more; go to bit.ly/newsinnovate
