Setup

Required libraries

pandas
numpy
matplotlib
sklearn

Running code

All scripts are in the src/ directory, and scripts with a number in the filename should be executed in that order.

1-just-downtown-vancouver.py

Args: input-path, output-dir
Returns: JSON file with only points that are roughly within Downtown Vancouver

$ python3 1-just-downtown-vancouver.py ../input/vancouver.json ../output/

2-interesting-heuristic.py

Args: input-path, output-dir
Returns: JSON file with the same points, but with an added interesting_heuristic column (normalized to [0,1])

$ python3 2-interesting-heuristic.py ../output/downtown-van.json ../output/

3-clustering.py

Args: input-path, output-dir
Returns: JSON file with the same points, but with an added location_cluster column (computed with KMeans clustering). If a point's heuristic is too low, location_cluster is null

$ python3 3-clustering.py ../output/heuristic.json ../output/

4-choose-places-to-visit.py

Args: input-path, output-dir
Returns: JSON file with points to visit (decided based on a combination of cluster and heuristic)

$ python3 4-choose-places-to-visit.py ../output/clusters.json ../output/

5-create-path.py

Args: input-path, output-dir
Returns: JSON file with a path through the points, ordered in a way that minimizes walking distance (based on a Haversine path through the points). Also prints path details and estimated total walking distance and time

$ python3 5-create-path.py ../output/places-to-visit.json ../output/

plot-on-map.py

Args: input-path, path-path (only if using -p), map-path, output-dir
Flags (optional): -h (heuristic heatmap), -c (clusters), -p (haversine path)
Returns: PNG image of data points plotted on the provided map. If no map is given, only prints the bounding box (coordinates of rectangular area containing all points), which can be used to download an appropriate map from openstreetmap.org.

$ python3 plot-on-map.py ../output/downtown-van.json  # bounding box
$ python3 plot-on-map.py ../output/downtown-van.json ../input/downtown-van.png ../output/  # scatterplot
$ python3 plot-on-map.py ../output/heuristic.json ../input/downtown-van.png ../output/ -h
$ python3 plot-on-map.py ../output/clusters.json ../input/downtown-van.png ../output/ -c
$ python3 plot-on-map.py ../output/clusters.json ../output/path.json ../input/downtown-van.png ../output/ -p