While enthusiastic citizen scientists record great numbers of observations and photographic evidence of species in networks such as iNaturalist, their numbers pale in comparison to the multitudes of images shared on generic social media platforms, like Flickr. Although the majority of such images have little relevance to biodiversity, being able to extract the few that does, could be a cheap and easy way of complementing existing biodiversity information.
Following this idea, researchers searched Flickr for geotagged images of two species of bees and flowering plants in Australia and used Google’s reverse image search to validate the images and exclude false positives. The locations of the images were overlaid on a map with occurrences of the same species obtained via GBIF and the Atlas of Living Australia.
The produced maps showed a general overlap between Flickr-derived images and the range of the GBIF-mediated occurrences, and if used properly, this big-data methodology could provide an inexpensive and abundant source of occurrence data.