Open BeSt Addresses

The project

In June 2019, FOD BOSA DT published the BeSt data set, a collection of XML files in which all addresses of Belgium are displayed with their respective geolocation. This data offers many possible applications, such as checking addresses or creating applications with geocoding. To make the XML data accessible for other applications, we created Open Source tools that enable smooth conversions of data. The converted dataset has been added to openaddresses.io, a global open address file that is already used by many applications. With this, we were able to set up a Geocoder service in no time.

The tools

To promote the use of the BeSt Addresses dataset we created a variety of Python tools to convert, filter and use the data. These tools are bundled in a Dockerfile, for which the installation instructions are provided on our Github repository.

With these tools we can easily...

Download the zipped dataset of the addresses. Extract the xml files to a directory. Convert to the csv format or other formats such as GeoJSON and Shapefiles. Filter the file on postalcode, region and bounding box. Use the addresses to match and complete other address datasets.

The geocoder

We managed to get the address data of Belgium uploaded into openaddresses.io, which is a global address collection. This dataset then can power many other applications, one of which is a geocoder

We set up a pelias geocoder without much trouble and before long we were providing geocoding services to other oSoc19 teams. The process to setup a geocoder using the Belgian address data is documented in our Github repository.

The notebooks

With the address data converted by our tools it became much easier to explore it. This allowed us to make a collection of Jupyter notebooks using the address data in fun and interesting ways