OpenFlights has always had a very liberal policy when it comes to entering planes for your flights: be it Boeing 737 or Santa’s Sleigh, anything goes, just type in the box and it’ll be accepted. Over the years, this has led to over 70,000 different names being entered, most reasonable if occasionally misspelled (there are 28 different Beoings in there). However, since OpenFlights will also create new planes from CSV imports, this has led to a lot of completely random cruft including large chunks of raw XML and one particularly misguided user uploading over 10,000 strings of gibberish. And because OpenFlights used to match against the whole lot when typing in planes, this made it harder and harder to find the needles in the haystack.
So I’m glad to announce that the plane autocomplete results have been given a long-overdue reworking. There is now a canonical list of 160-odd planes that will always show first, so entering 747 will default to Boeing 747, and common three-letter IATA codes like 772 for Boeing 777-200 are also understood. For everything else, results are ordered by frequency of usage, so entering Tupolev now defaults to the fairly common Tu-154 (~2600 flights), not the hypersonic Tu-144 “Concordski” which hasn’t flown since 1978 and which only 3 Openflyers claim to have boarded.
What’s more, the canonical plane data is now released as open source, so you can use it to demystify the IATA and ICAO codes often found in schedules and more.
This is only the first stage in a data scrub that will next move onto airlines and finally finish off with airports, which were already given a quiet but incomplete reboot last year. Stay tuned for more!