PyData Berlin Geocoding / Mapping Special
Image source: Mircea Iancu

PyData Berlin Geocoding / Mapping Special

We kicked off November with a PyData Berlin meetup. Pizza, drinks and festive Spekulatius, along with three great speakers of course, guaranteed a great turn out and we filled up all the seats we had available. Thank you to all the guests and of course to the speakers. Below you’ll find a few pictures from the event and the details of the talks.

Geocoding at Comtravo

Comtravo is a business travel agency with a focus on making business travel radically simple. The Data Science team at Comtravo extracts travel booking requests from customer emails. One part of this process is identifying and resolving location mentions to specific locations on the planet. For instance, flight requests need to have an origin airport and a destination airport, both resolved to an IATA airport code, but hotel requests require resolving hotel names to unique IDs. Finding out which one of the Motel One Berlin hotels was meant requires more than a simple search.

Predicting the Spatial Distribution of Ridepooling Demand with Machine Learning

Door2door’s mission is to reduce urban traffic by enabling cities and communities to provide a Ridepooling service as an alternative to private cars. The success of this service depends a lot on operating it at the right time in the right area. In the presentation, Sabine Kaiser will show how to use machine learning to predict the spatial distribution Ridepooling demand in urban areas based on a variety of openly available geographic features.

Exploring HERE Location APIs

This workshop style presentation will outline essential RESTful APIs offered by HERE.com. Dinu Gherman (senior engineer at HERE Technologies) will demonstrate many examples for these APIs in a live Jupyter notebook Python environment, and include inputs from the audience to build a small, but useful application on the spot. Dinu will cover at least the following APIs: geocoding and reverse geocoding, places, map tiles and map images, routing, isolines, traffic, transit, places, and map-matching.

GeoCoding at Comtravo

Predicting the Spatial Distribution of Ridepooling Demand with Machine Learning

Exploring HERE Location APIs


Written by

Dr. Matti Lyra

Research Engineer in NLP

Matti has been an active member of the data science community in Berlin, he has led the PyData Berlin committee since 2016. Matti holds a PhD in Natural Language Processing from the University of Sussex. His thesis explored using topical bias in large text corpora to build better ensemble models for text classification. Beyond his work at Comtravo he's interested in language and linguistics, the philosophy of AI, cycling, photography and coffee.


Updated