This project includes a collection of Python scripts to download Purple Air air quality (particulate matter) data from the thingspeak.com API for a set of sensors that are within a specified geographic region (minimum/maximum latitude/longitude).

The scripts include methods to download data from the thingspeak API, merge and augment the data and then perform various spatial and temporal analyses along with visualizations to be able to comprehend air quality across space and time.

The scripts build a table of all the registered PA sensors on the network in a sqlite database then use that table as a source for the sensors of interest, where they are and when they last reported.

For now, this repository includes only the download scripts as I am still working on the analytic elements.

This script reads from the master list of Purple Air Sensors JSON file (https://www.purpleair.com/json) to identify all the sensors that have been registered to update the local sqlite database.

For each sensor, it checks whether it has seen it and if so, update last_seen_date used in later scripts. If it hasn't seen it, it then creates a new record.

This script also has a call to the Google Maps reverse geocoding API that can be used to also collect the City / State for each sensor. I used this for other projects.

Download the raw purple air data from thingspeak within a geographic area of interest provided as arguments to the script (see below). We then loop over each sensor in the sqlite database that are within that bounding box and download daily raw data for each sensor from it's installation date through today. This data is currently downloaded as CSV into the data/thingspeak folder

python download_raw_purple_air_readings.py 33.994905 34.057430 -118.510778 -118.427848 72

The previous step can create hundreds of CSV files. This simple procedure merges them into one single file.

Coming soon