Critical Mass Essen january 2019 tour

Data Visualization

At critical mass events bicylists meet and drive through a city as a convoy. This creates awareness for cycling as a mode of transport. I logged the january tour and turned the data into an animated map:

Due to rain the tour was shorter than usual, taking about 83 minutes. With a distance of 15,1 km covered the average speed was only about 10,9 km/h. The convoy drives at a leisurely pace and slows down at many traffic lights in the inner city area.

The first kilometer seems a bit short. This was due to circling at the start of the tour and circling in two roundabouts. These movements only become visible when the animation is played at small time intervals. A fast movement can be seen at about 20:07. There the convoy drove downhill through a tunnel. The right/left orientation of the bicycle icon is based on the logged bearing data, i.e. the direction of travel. Some flickers can be seen in this data, especially during stops.

The orthophotography recolored to a monochrome blue makes the city look more densely covered than in the other maps I‘ve seen. This is of course mainly due to the green color being lost, but also due to other factors. I’ve noticed that cemeteries, garden colonies and parks are mostly drawn green in standard maps, while they appear similar to urban fabric in satellite photography. This reminds me of using less processed maps and more raw satellite imagery when plausibility checking geographic data.

The animation was produced in the following workflow. I logged the tour with the GPSLogger app. I used a self-written python script to convert the time sequence of points as a CSV file into a time sequence of lines summarizing the tour up to that point as a geojson file. I set up the tour data in the QGIS Time Manager and exported the frames as PNG images. I loaded the images as layers in Gimp, applied the gif-animation filter for optimization and exported it as an animatedGIF. The GIF file I additionally converted to a MP4 video file using Shotcut.

Some technical details are noteworthy. Especially superimposing the labels and icons in QGIS required some tweaking. I downloaded the bicycle icon as a SVG file from wikimedia commons and constructed a right/left version of it. The logged data contains a bearing field (direction of travel). Using rule-based labeling, for a bearing up to 180 degrees a right-driving bicycle is shown, for larger bearings a left-driving bicycle is shown. QGIS always places labels above icons. This can be worked around by making the icon part of the label by using it as a background image in the label settings. The kilometer markings were set to a low priority and to alway draw to keep them in the background. Dark blue halos were used for all labels to make them better visible. For normal symbols, the outer shadow effect can be used to emulate halos, but for icons used as label backgrounds only the weak shadow functionality is available. Thus I added the icon halos directly into the SVG files using Inkscape.

The orthophotographs I downloaded from the open geodata portal of Nordrhein-Westfalen. The JP2 files (JPEG 2000) can be directly loaded as georeferenced data into QGIS. To get a monochrome image I converted the image to greyscale and colored it blue in the layer settings.