Healthy Hives: Cloud Analytics Helps Save the World’s Bee Population

Machine Learning Case Study

Oracle Analytics Cloud is the foundation for an ambitious project aimed at finding the cause and cures for honeybee population decline.

An alarming decline in the global bee population has been making headlines recently and for good reason: An estimated 77% of the world food supply depends on pollinators, meaning that the human population—and up to $600 billion worth of agricultural activity—is at risk.

Combating the risk starts with collecting, consolidating, and analyzing massive volumes of data across the world. Oracle is working with the World Bee Project to bring the necessary cloud services to bear on this issue. The Global Hive Network is an initiative to collect billions of individual data points from around the world and analyze them to understand the honeybee population’s overall health and its relationship with environments, weather patterns, forage, diseases, parasites, predator species, and pesticides. The knowledge is being shared with farmers, beekeepers, and researchers around the world.

Thousands of Species, Millions of Hives

Crops requiring pollination include fruit, vegetables, seeds, nuts, and oils, many of which are important dietary sources of vitamins and minerals. Several crops, such as coffee and cocoa, also are an important source of income in developing countries.

With more than 20,000 species of wild bees spread across the globe, understanding their behavior, food sources, predators, and other threats means recording and analyzing data with significant local and regional variations. In the U.S. alone, more than a million hives are relocated to California’s Central Valley from across the country every February to pollinate the region’s almond crop.

Thanks to advances in the Internet of Things (IoT), individual beekeepers have been able to monitor the health of their hives. Sensors record the hive’s weight, internal temperature, humidity, and even sound levels. For example, acoustical monitoring has helped beekeepers prevent “swarming,” when about half a hive’s population departs to a new location.

Swarming makes honey bees highly vulnerable to starvation or predators, leading to population loss. Experienced beekeepers can detect a faintly audible “warble” that indicates the hive is close to swarming and move it to another area of the crop field before it does so. Acoustical sensors are not only better than humans in detecting the warble, but they can also send an alert remotely so that the beekeeper doesn’t have to physically monitor each hive. (The technique may well prove beneficial in other areas beyond beekeeping, such as preventative maintenance of vehicles and equipment.)

The Value of the Cloud

Acoustical monitoring for swarming is just one piece of the puzzle, but it’s a dramatic example of the volume, velocity, and data variety capabilities that cloud computing brings to the table.

One hour of acoustical data from four hives generates more data than a Tesla in the same time period. On-premise infrastructure would have to be built to accommodate potential peak data capacity, consuming resources and wasting money. The cloud’s elastic data storage capacity allows for data collection and storage when needed with the ability to scale back when it’s not.

In addition, machine learning in the cloud can build predictive models based on data aggregated across thousands of hives to better anticipate swarming. Similar processes apply to other data streams that indicate hive health, such as a hive’s weight and its internal temperature and humidity levels. Based on observations of healthy hives, the cloud solution can generate alerts to hive owners when trouble arises and give specific advice as to possible underlying issues. It can also produce data visualization to help researchers glean additional insights into bee population health.

All this information is available to beekeepers, farmers, and researchers in real time and on mobile devices. They can interrogate the data via chatbot to ask specific questions about a hive’s health and receive alerts when anomalies arise so they can manage by exception.

Other IoT and cloud technologies are proving highly useful as well. Cameras can report the presence of predators such as wasps, and blockchain assists in assuring that the honey that arrives on retail shelves is authentic—important because up to 76% of honey sold has been over-processed or adulterated with cheap ingredients, including harmful ones. In addition, cloud analytics can help reduce harmful pesticide use by pinpointing the exact location of infestation.

The World Bee Project Global Hive Network is built on the Oracle Cloud using Oracle Analytics Cloud. Together, these solutions provide access to the massive data quantities generated and interpret that data aided by artificial intelligence, machine learning, and sophisticated visualization, all with the flexibility to scale as needed.

Contributed by: Rich Clayton, Vice President of Analytics Product Management, Oracle. He has a passion for helping companies transform their organization with data by researching market and technology trends, and engaging customers and partners.  He presents at analytics forums around the world on Oracle’s strategy and the future of analytics, machine learning and artificial intelligence. He teaches analytics at University of Oregon and is the executive sponsor for the annual analytics challenge at CalPoly.  Mr. Clayton earned his bachelor’s degree in accounting from Loras College.

Sign up for the free insideAI News newsletter.