PyData 2019

Attila Balázs, Principal Consultant @Sword Venture gives us his view of PyData London 2019:

PyData was great. I enjoyed seeing how data science is applied in other industries and met a lot of interesting people.

My favorite talk was about the process of Active Learning delivered by Jan Freyberg from Element AI. Active learning is an iterative process to involve domain experts into the ML training workflow, it is especially useful when you have a small labelled dataset and large unlabelled datasets.

Connor Tann from BP gave a great talk about Geoscience and how data science is helping to interpret the subsurface more efficiently. BP was one of the main sponsors of the event. There were lot’s of colleagues from BP at the event, lots of familiar faces.

Another highlight was seeing the platform Man AHL created for dealing with 1 to 3 billion daily stock market signals. They have built their own time series database, which has been open sourced and use Apache Kafka heavily for real time data processing.

Octopus Energy also delivered a great presentation on forecasting their customers energy demands. They created and open sourced a library for timeseries data and showed us their approach using Keras and Tensorflow.

A very exciting event… I am looking forward to attending next year!