Bumblebee observations & flower image challenge


We have a data set that contains 18,000 observations with pictures of bumblebees, collected in the last 15 year on Waarneming.nl. This challenge however is not targeted at the bumblebees in the pictures, but at the many unidentified flowers in the background.

We have already enriched the data set with machine prediction of “phenology” (flowering, vegetative, “no plant”), based on manual annotations (~2,500) and predictions (18,000) using the NOUS platform (https://cosmonio.com/) and a first identification of the flower (plant) species using the Naturalis’ Nature Identification API (https://identify.biodiversityanalysis.nl/).

The challenge for this hackathon is to discover unknown relations in the dataset or with other datasets, for example preference of bumblebee for plant species, relation between visits to flower and time of day/season or relations with GIS datasets.

About the Big Data for Bees Hackathon
Who is Who at the Big Data for Bees Hackathon

Challenge and data can be found here: https://docs.google.com/presentation/d/1yPuMHXWiM1zo5DzTCC-M7t8CnPASWynfJ-SqVDhDGDU/edit?usp=sharing


Hello, how is it going? Ready to pitch to the jury? :slight_smile:


Our fantastic fabulous presentation!


Hi Hommeles,

I had to run home at 14.00 because my kids were home alone…

Heared from Sonne both our teams won the first price!

Great work you have and done, am sure we’ll meet soon and get some Bee-effect going sunny: