/firestation-clustering

Primary LanguagePython

Firestation Clustering

How to run

  1. Install https://python-poetry.org/docs/
  2. Copy config-example.yaml to config.yaml
  3. Insert api-key in config.yaml
  4. poetry install
  5. poetry run x

ToDo

  • Switch everything to mapbox
  • Probabilistic fire spawning
  • Log all relevant data in every iteration
    • Firestation locations
    • Fire locations
    • Driving times /euclid distance
  • Add kmeans_driving_time results output
  • Run experiments
  • Calculate final average driving time for all experiments

Experiments

  • euclid uniform
  • euclid weighted probabilities
  • haversine uniform
  • haversine weighted probabilities
  • driving_time uniform
  • driving_time weighted probabilities

Let's assume we have 1000 fires for the final metric. How many matrix elements do we need? -> 1000 * 4 * 6 = 24000

For driving_time training we need -> 2 * 20 * 400 * 4 = 64000

Total matrix elements: 88000