The Data Game

Team member checking a camera-trap

So we all know there are millions of images on snapshot Serengeti and that it is us citizen scientists who do all the work classifying them. The scientists can then get on with the task of figuring out what’s going on out there in the animal kingdom, hopefully in time to save some of it from our own destructive nature.

But… have you spared much thought as to how the images go from over 200 individual camera-traps dotted around the Serengeti to the Zooniverse portal  in a state for us to start our work.

Firstly the SD cards have to be collected from the cameras and as this is an ongoing study replaced with fresh SD cards. This is done about every 6 to 8 weeks. A camera traps batteries can actually go on performing far longer than this but as the field conditions can be tough you never know when a camera may malfunction. This time frame is a good balance between not ending up with months worth of gaps in the data and not spending every minute in the field changing cards.

The team are able to check about 6 to 10 sites a day so with 225 cameras in play it takes around a month just to get to each site. Mostly the cameras are snapping away happily but there are always some that have had encounters with elephants or hyena but actually some of the most destructive critters can be bugs, they like to make nests of the camera boxes. As well as checking the cameras themselves the sites need to be cleared of any interfering foliage, we all know how frustrating a stray grass blade can be.

Snake camping out inside the camera-trap

So with a hard drive full of all the data it then has to wait for a visiting field researcher to hand carry it back to the University of Minnesota, USA. It means the data is only received every 6 months or so but it is far safer than trusting the post. Once safely received it is up to Meredith to start the painstaking work of extracting the date time stamps. As sometimes happens there are glitches and she has to fix this by figuring out when the camera went off line or when capture events got stuck together. She says it is much like detective work. The images are then assigned codes and stored on the Minnesota Supercomputer Institute (MSI) servers.

Once it is all cleaned up and backed up it is sent to the Zooniverse team who then format it for their system giving new identifiers to each image. Finally it is ready for release to all the thousands of classifiers out there to get to work on.

So as you can see it really is a team effort and a massive under taking. It is no good collecting tonnes of data if there is no one with the time to do anything with it. I will take this opportunity again to thank you for all your help with the project. Keep up the good work.