FOSDEM is a free event for software developers to meet, share ideas and collaborate, organised every year in Brussels. It is very very geeky. You find all kinds of people working on all kinds of open source software, from Linux distributions to Libre office, from databases to programming languages, from open source micro controllers to open source big data platforms. Alessio Fattorini called Fosdem the Woodstock for geeks on Twitter.
We go every year, as there is one dev room that is very interesting to us, the geospatial dev room. We use lots of open source geo related software, such as GDAL, Geopandas and POSTGIS. And for the first time, we submitted a talk and were selected to speak!
First things first, here you can find Mathieu’s talk about point cloud processing. Most of it is also covered in this previous blog post.
We use GDAL all the time, and man it gave us nightmares. In this talk by Jeremy Mayeres you get an introduction to GDAL, plus some interesting links. We learned that there is a Python wrapper around it, which avoids you to use subprocesses all the time, which is very nice!
One of the main contributors pandas and geopandas, Joris Van den Bossche, gave a talk about Geopandas. Geopandas allows you to make spatial dataframes in Python, which are dataframes with a geometry column. We learned that you can use this column to make spatial joins, which is a handy feature.
Deep learning is everywhere. This talk by Olivier Courtin gives a introduction about pixel classification with deep learning of map images. There are a few Kaggle competitions organised to get you started in this geo data science field.
And last but not least, a big thank you to Johan Van de Wauw for organising the talks in the geospatial dev room.
We were inspired by all the things we learned and especially by this tweet to make a little map, representing the travel time to the city center. Give us a should if you want to take a look at the notebook (the code is a bit too messy to publish it right away).
See you next year!