PyData Amsterdam 2017

I am attending this weekend the PyData Conference in Amsterdam.

The talks I enjoyed the most after two very busy days:

i) Diagnosing Machine Learning Models by This talk was very engaging and very simple to follow and understand. It touched on how to identify multicoliniarity in models and how to reduce/eliminate it. Click here.

ii) The second, Deep Learning at on Wide & Deep Learning for Recommender Systems (published by Google). Click here.

iii) The third, Deploying Python models to production. A very nice and inspirational talk given by a data engineer this time who presented a unified workflow for automating the deployment of models into production environments. Click here.

Next to these talks there were plenty of other very cool presentations. However, most of them were quite theoretical or went quite deep into certain theoretical aspects which made them harder to grasp or understand their utility in practice. I very much enjoyed the work done to monitor the behavior of orangutans sent back into the wild. This has been achieved by using a drone flying over the jungle and picking up signals from these animals, making identification and monitoring quick and easy. For more info click here. #coolstuff

That’s it for this year’s PyData conference in Amsterdam. Very cool, very inspirational and certainly very useful.