Hi,

Inspired by @dehbaiyor I would also like to share my solution in exchange for getting a glance at yours .... :-)

My solution made the 34th place with a score of 3.75.

My solution uses the Uber Movement data and resulted in 13 features. My final model was a random forest. I evaluate a wide range of models and frameworks, thereunder frameworks for automatic feature selection and autoML. I put effort in feature selection and looked at feature importance, correlation, trend correlation, permutation importance and dentrogram. All in one compact notebook, cleaned up and freshly annotated.

Please request the notebook by sending me an email to tf@computas.com, with your solution attached or linked.

With kind regards,

-Ture