Detecting Bee Hive Behavioral Changes Through Frequency And Signal Analysis Of Audio Files

Honey bees (Apis mellifera) are among the most important organisms in the world. The pollination they provide is crucial to the survival of many economically-significant plants, but since the late 1990s, beekeepers around the world have reported significant declines in domestic honey bee populations. In response, researchers have developed new methods of analysis to track the health of their hives, and one such method involves analyzing the sounds produced by a colony. Honey bees communicate through sound signals that occur within specific frequency ranges and signify different behaviors, and by determining the signals that are being produced, the behavior of a colony can be better understood. As a part of this research, software was developed to plot the spectra of audio recordings from domestic hives. For each individual audio recording analyzed, the software computed the average magnitude throughout multiple frequency ranges along with the change in average magnitude between specific frequency ranges and exported this data to a spreadsheet for further analysis. These values were used individually and together to quickly identify audio recordings that stood out from a quantitative perspective, and these recordings were then individually analyzed in order to locate instances of strange honey bee behavior.