The project is organized into 4 activities reflecting the collaborative effort involving the Computer Science, Data Science and Biology disciplines:
Activity 1: Data collection and annotation
Activity 2: Video analysis
Activity 3: Data Mining
Activity 4: Biological interpretation

These activities correspond to the following research objectives:

Computer Science objectives (Activity 2): Detect, track and recognize behaviors such as foraging and fanning at individual level from video acquired continuously during months of observation
– Automatize video annotation of bee behavior to enable large-scale data analysis
– Provide tools to assist expert end-users in training new behavior detectors efficiently.
– Scale-out the computations on High-Performance Computing (HPC) cluster to enable high throughput analysis, while providing remote access to let domain experts provide input and visualize the results

Data Science objectives (Activity 3): Find patterns in these behaviors and their relation with a variety of other data collected about the bee environment
– Develop visualization tools adapted to complex activity patterns that involve multiple parameters and time information.
– Develop data analysis suitable for the identification of clusters and correlation in such data

Biology objectives (Activity 1 and 4): Gain new insights in the effects of individual variation in behavior with colony performance
– Collect and generate massive amounts of behavior data from bees in their natural habitat
– Determine the basis of individual differences in the timing of foraging and fanning behavior.
– Correlate circadian parameters with the timing of foraging and fanning behavior, locomotor activity and colony performance