Spatial: Spatial Data Projects (2025)

I extended my exisiting data science/engineering knowlege to spatial information and its analysis and presentation. here are some snapshot of my journey and ongoing projects:

Berlin Bike Risk

Do bike infrastructure projects reduce accidents with injury? For this, contrast history data on accidents between bikes and motorized vehicles; 2018/19 vs 2020/21/22. We chose to consider Berlin, and a neighbourhood close to the wall with a lot of demographic change and thus potential infrastructure debt in the last decades, Neukoelln.

Topics

  • Geoinformation
  • Cartography, Visualization
  • Spatial Data Science

Methods

  • ArcGIS, QGIS
  • Data Acquisition, Management/Storage, Processing and Analysis (2D/3D)
  • Web Maps
  • OSM Query
  • LIDAR, Raster Processing

Historic context: Berlin wall and the corresponding Mauerweg bike paths (Historic aerial picture overlaid with features from Berlin Open Data / GIS)

Hotspot analysis pre/post infrastructure projects

We can see that there are several projects addressing the high accident count on Hermannstr., and a currently undertaken remodel of the first stretch from Hermannplatz south. On Sonnenallee, there many accidents, so there seems to be a new alternative bike route connection north-east in planning.

  • Basemap: OSM Blueprint WMS from ESRI, 2024 final edition
  • Strassennetz: Berlin from Living Atlas; reclassified StEP-Klasse to 0..50 Numeric values with Python Notebook; definition query to only minor roads (>20) symbolized as proportional lines with corresponding color.
  • Bicycle infrastructure projects: Scraped from infraVelo API, filtered for bike paths and translated the nested JSON to GeoJSON/shape file that includes the details
  • Bike Paths (Berlin Open Data)
  • Accidents: 2018 / 2019 / 2020 / 2021 from Berlin Open Data
    • Melted into one table and augmented with labels from reference PDF; filtered by accidents with bikes
    • Symbolized as points, injury degrees as size and as heatmap

Cartography for the Community Garden

Working on a new map for the community garden, based on the OpenStreetMap data (yet to be extended with more details)

For the basemap:

  • Since the DGM1 was so low in resolution, I created a DEM from the 3D LIDAR scans of NRW by interpolating between the ground-classified points.
  • For the surround map, I rendered the NDVI Normalized Difference Vegetation Index of the Sentinel-2 data on a green shade to provide an impression of the surface use (fields, forests)
  • For the detailed map, I symbolized by OSM classifications.

Orthophoto 2023 (CC BY 4.0, Stadt Bielefeld - Amt für Geoinformation und Kataster)

Classified LIDAR data, with experimental 3D-lines for the internal footways.

Comparison of the DGM1 (red contour) with the new DEM (blue contour)

Shaded relief of the new DEM

Hot in Amsterdam

Lets go to Amsterdam on the hottest day of the year, to escape the heat a bit, they said. Lets make a quick said; 1-hour hit-n-run medley in ArcGIS Online of the AMS Open Geodata possibly relevant. Sadly, no (more) heat map available, just noise (Geluiden). It helped finding some point of interest. Also checked out the netherlands in 3D to see where the shadows would be. Why can’t jam with layers like this in a normal navigation app like Organic Maps?

City Scale

Vondelpark Detail

Legend

Speed and Silence in the Green Hell

I visited the Nordschleife with a linse for the Rad am Ring 24h bike race, who gifted a GPX track of one of our spectator laps during the race. So slow uphill, so fast downhill, so mesmerizing the silent soundtrack of perfectly oiled road bike gears in the green hell. Quick visualization of the speed / position on the beautiful hillshade of the RLP geo portal.

Pink is faster, blue is slower