Clean Air Compass - Mapping API - Serverless Version

An API to handle location requests from the Clean Air Compass frontend and return an interpolated map of air pollution (PM 2.5) in the form of a list of georeferenced polygons.

Details

The API conducts the following steps:

• Parses the location request to see if it is for a city or a US postal code
• Makes a request to the LocationIQ API and get back a bounding box of lat / lon coordinates
• Makes a request to the Purple Air API to get back the list of sensors within the bounding box as well as their associated air pollution data
• Creates a spatial data frame from the sensor API response object using shapely
• Creates a grid of polygons based on the spatial extent of the collection of sensor points and estimates the PM 2.5 pollution at each grid cell using the K-nearest neigbors (KNN) regressor method.
• Returns the grid of polygons to the frontend as a GeoJSON object

How to run

1. Clone the repo
2. Install all dependencies by running pip install -r requirements.txt in the console
3. Create your own environment file where you set the values for your own Location IQ API key (LOC_IQ_KEY) and Purple Air API key (API_KEY) respectively.
4. Run the API server locally using uvicorn main:app --reload