Create and customise a heatmap of your bike rides.
My heatmap of 4 years of Deliveroo cycling with no background layer.
My heatmap of 4 years of Deliveroo cycling with building and water layer behind.
The resulting getagged .tif file can be imported into ArcGIS or QGIS for further analysis or rendering opportunities.
Other scales are also possible. Below is the heatmap of my recreational rides in the vicinity of Bristol.
The python packages can be installed using pip install -r requirements.txt.
The GPX files should be placed in one folder. Each GPX file should contain only one GPS track. The frequency of samples in the files should be constant. For best results the sampling frequency should be 1Hz (one sample per second).
The example folder is an example folder with a few GPX files to demonstrate the command line arguments of the script.
The required argument -folder or --f sets the folder which contains the GPX files. It searches all subfolders for GPX files in a recursive fashion too.
The script can be run by simply writing:
python3 draw.py -f example
The extent of the map can be defined by the --area or -a argument. It expects a formatted string in the form "xmin,ymin,xmax,ymax" where xmin is the minimal longitude, ymin is the minimal latitude, xmax is the maximal longitude, ymax is the maximal latitude. If any of the four number is not given, it is inferred from the data itself. For example if we want to crop only the bottom we should run the following command:
python3 draw.py -f example -a ,51.464753,,
If we want to crop all four sides then run:
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004"
We need the surrounding quotation signs and the extra space because the starting coordinate is negative.
The algorithm divides the geographical area into a grid, and places all coordinate samples into the grid cells and counts the number of points in each cell. We can change the resolution of the underlying grid by the argument --bins or -b (default = 1200). We want to be sure that we get this parameter right because both too large or too small resolution might be unsatisfactory. An example use:
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300
The next argument is --limit or -l (default = 30). When a place is visited way too often than other places it makes all other places dark/invisible. By setting --limit lower we can add more colour to less visited places, by increasing it we can emphasise the most visited places. For example:
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300 -l 150
The argument --colormap or -c (default = inferno) can define the color-coding of the heatmap. We use matplotlib colourmaps. These colormaps can be used in the argument: https://matplotlib.org/stable/tutorials/colors/colormaps.html. For example another coloring:
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300 -l 150 -c turbo
Places visited very few times can seem to disappear from the image. By the --add_min or -m (default = 0) argument we can add more emphasis to these places, which means we compress the colours a little.
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300 -l 150 -m 50
There is an opportunity to smoothen the image. This can be achieved by the -smoothing or -s (default = 7) argument. It expects a non-negative number. 0 means that it does not perform smoothening. In the argument, the number n means that it will perform smoothening with a kernel of size 2n+1.
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300 -l 150 -s 0
Places where we stopped during our activities can be emphasised by the argument --emph_stops or -e (default = 1) argument. It expects a postive number, and the bigger this number is the more radical the effect will become.
python3 draw.py -f example -a " -2.602665,51.444753,-2.592001,51.453004" -b 300 -l 150 -e 8
From the website https://download.geofabrik.de/europe/great-britain/england.html and other websites we can download shapefiles. Put them in the XML file and render it. Make sure the heatmap .tif file is interpolated either in bilinear or bicubic manner for best results.