Social networks of honey bees can be analyzed to learn more about various topics like the characterization of the swarm, work tasks by age groups, transmission processes or even disease dynamics. Social netowrks can be build through contact events. One important contact event of such networks can be transmission of fluids - trophallaxis. Finding trophallactic contacts manually is time consuming. Automatic detection or at least an assisting system would be useful to be able to classify large data sets. This work attempts to develop a system to detect trophallactic contacts by filtering spatio-temporal data. The results are then discussed.