Repo for paper Incorporating Dynamic Flight Network in SEIR to Model Mobility between Populations.

See requirements.txt.

This project leverages pyflightdata API, a wrapper around flightradar24, to fetch flights data.

.
|-- datasets
|   |-- airports
|   |-- edge_lists
|   |-- historic_flights
|   |-- unique_flights
|-- graphs
|   |-- covid
|   |-- flights
|-- logs
|-- src
|   |-- data
|   |-- model
|-- states

1. Fetch all countries: python src/data/fetch_countries.py
2. Fetch all airports of all country: python src/data/fetch_airports.py
3. Collect a list of flights departing and arriving a the given country: python src/data/fetch_unique_flights.py --countries Canada Note: due to limitations of pyflightdata, fetch_unique_flights.py can only fetch live flights. In order to obtain a comprehensive list of flights to and from Canada, one must run this script repeatly over a long period of time.
4. Fetches the historic flights of a given country: python src/data/fetch_historic_flights.py --countries Canada Note: by default, pyflightdata only returns historic data for the past 7 days. In order to obtain 90/360 days of past flights, one must subscribe to the Silver/Gold plan on flightradar24 and login programmatically with f.login("username", "password").
5. Download owid-covid-data.csv from Total COVID-19 Tests Performed by Country and place it under datasets folder.
6. Download can_covid19.csv from Coronavirus disease (COVID-19): Outbreak update and place it under datasets folder.
7. Download reported_flights_stats.csv from Domestic and international Itinerant movements and place it under datasets folder.
8. Download canada_population_by_prov.csv and place it under datasets folder.

1. Construct flight network edge list for Canada: python src/data/construct_edge_list.py --country=Canada followed by python src/data/process_flights_data.py
2. Construct flight network edge list for provinces: python src/data/construct_edge_list.py --use_country --country=Canada followed by python src/data/process_flights_data.py --use_country
3. To process test positive rates for the world, run src/data/process_positive_rate_world.ipynb to produce test-rate-processed.csv
4. To process test positive rates for provinces, run src/data/process_positive_rate_can.ipynb to produce can_covid19_processed.csv
5. To generate interpolated test positive rates, run src/data/interpolate_positive_rate.ipynb to produce interpolated_positive_rates.csv
6. To generate data used for experiments, run src/data/flight-seir-Canada.ipynb. The generated data will be saved in the states folder.

Run src/model/early_time_prediction.ipynb or src/model/reopen_simulation.ipynb.