discussion / Biologging  / 17 February 2021

Machine learning for animal tracking Kaggle competition?

Hello everyone, 

My name is Kieran Garvey and I work with Cambridge Spark - a specialist AI and Data Science education provider. https://cambridgespark.com/

We are looking to launch a Kaggle competition in June or July which is basically a place to run competitions on datasets to enable people to put forward their best machine learning/AI models. 

We thought it could be really interesting if we could work on an animal tracking dataset using historical data to predict animal movements/behaviours.

The key challenge here is that part of the dataset must not be in the public domain as this would enable people to cheat and look online to help them build the best models.

We would likely have sponsorship available for this dataset so please do get in touch if you think this is something that you would like to be involved with or know of a dataset that this could be applied to. I hope the models or insights created through this competition could somehow be useful in practice which would be awesome!

Thanks so much,

 

Kieran

[email protected] 




What kind of data are you looking for? Like just lat/long coordinates, GPS points, .shp files, etc.? I have a small amount of GPS data from focal follows of lemurs (+ behavioral context/individual ID/subgroup, etc.) over a 3 month period. It's not published as I'm only using it in conjunction with a vocal dataset I'm working with. But I imagine you're probably after much larger datasets?