Airport POIs (BETA)
This is a BETA version for testing Airport POIs
I use a list of aiports and coordinates by lxndrblz/Airports, shared under the Creative Commons Attribution Share Alike 4.0 International license.
There is a commercial dataset of airport traffic, but it costs $6k. I considered buying it, but for now let’s create some reasonable inputation.
1. Collect annual passenger statistics for top 50 airports from Wikipedia.
2. There is a global free dataset of flight origins and departures from January and February 2020, the COVID-19 AirTraffic project, MIT license).
3. Using a ratio between (1) and (2) for the top 50 airports, I impute all other airports.
To then translate airport annual passenger statistics into NIMBY population, I use the default demand curve. I then choose a fixed % of annual airport passengers computed above, such that the railway pax usage would match that of a few known airport railway stations (I looked at large airports UK, FR, NL and took an average).
This approach has several drawbacks, including:
– The COVID-19 AirTraffic project coverage is quite poor, but not zero, for many non-OECD countries. This means you get very large airports in certain countries assigned very low population.
– China is not covered at all by the COVID-19 AirTraffic data, the only China data included is that of a handful of planes from China that land elsewhere with origin data populated.
– The data is from January and February 2020, which means several popular summer holiday destinations (eg the Azores etc) don’t see much traffic during that time.
– The imputations I use are very imprecise.
So I know the data is pretty poor, and for now see this as a test thing only.
