Our role: Data scientist consultancy firm
Target audience: Agricultural practitioners, Beekeepers/Apiarists
Bees are one of the most essential components of modern agriculture. Most plants need to be pollinated by animal visitors, especially insects. A wide variety of flowering plants, such as the everyday visible fruits like apples and blueberries, require managed pollinators to be able to ensure production. Honey bees are ideal for pollination because they are easy to move and manage. However, during the winter of 2006-2007, some beekeepers in America began to report unusually high losses of 30-90 percent of their hives. Colony loss has declined since then but is still a concern. Since the number of bee colonies is closely related to agricultural production, there is a need for agricultural decision-makers and practitioners to know the status of local bee colonies, especially the number, loss trend, and colony stressors.
To meet the needs of agricultural development, we decided to build a data visualization app with open data to ensure that agricultural practitioners can explore bee colony data visually. Our application will display the number of bee colonies in each state of the United States over the years and allow users to explore changes in colony numbers and colony stressors by filtering by state and time.
The honey bee colonies and stressors dataset were obtained from TidyTuesday who in turn obtained the raw data from the United States Department of Agriculture (USDA).
The USDA collected the time series data from a stratified sample of operations across the United States that responded as having bees on the Bee and Honey Inquiry, and from the National Agriculture Statistics Service list frame. Data was collected on a quarterly basis for operations with five or more colonies, and on an annual basis for operations with less than five colonies.
In our dashboard, we will visualize colony_max and colony_loss_pct
on a geographic map of the United States. The variables colony_max and
colony_loss_pct are the maximum number of colonies in a state for a
particular time period, and the number of lost colonies since the start
of the time period divided by colony_max, respectively. We will also
visualize colony_n, the number of colonies in a state, as a time
series plot. Our final visualization will be the percentage of colonies,
stress_pct, that are affected by different colony health stressors,
stressor, for a particular state over time as a stacked bar plot.
Our project, the Bee Colony Dashboard helps identify the rate and
possible causes of colony collapse disorder over quarterly periods in
the US. Some key questions that can be answered are:
-
Which state has the highest/lowest colony loss percentage over time?
-
What are possible causes of colony collapse disorder?
-
What specific problems do certain state’s colonies face over time?
Below is a usage scenario of our dashboard:
Buzzing Bee is a beekeeping company interested in researching colony loss across the US. Since loss rates are high, the beekeepers are under pressure to create new colonies to offset losses each year. The company wants to work with agricultural practitioners and farms to keep their bee colonies stable without the need to add new colonies each year.
John, is an apiarist working in Buzzing Bee and is assigned a project to identify states where colony loss percentage is very high and to study possible causes.
With the Bee Colony Dashboard, he will be able to identify bee colony losses over quarterly periods in the US. The geographic map visual illustrates how bee colony losses vary over each state for a single period in time, and apiarists/beekeepers can gain an idea of the local severity of colony stressors in different states. The time series plot visualizes the number of colonies over time for a single state and will inform apiarists/beekeepers of how problems affecting colonies increase/decrease over time. The third visualization, the stacked bar chart of stressors, gives insight on what specific problems a certain state’s colonies face over time, and can provide crucial information on what stressors need to be combatted to improve colony health.