Aircraft Registration Prefix Dataset
Flydenity is a callsign identification library to help match tail numbers or
callsigns to origin nations. The library is a python wrapper on top of a curated
dataset containing a set of regular expressions generated from the
International Telecommunications Union (ITU) International Call Sign prefixes.
The registration codes in this dataset are separated by country post The Paris
Convention in 1929. The database also contains a description of each codeset
with 2 and 3 letter ISO country codes following the ISO-3166 standard.
Flydenity is on PyPi, simply install it with PIP
pip3 install flydenity
To run, you can simply include it in your python library using the following
>>> from flydenity import Parser>>> parser = Parser()>>> parser.parse("AF1234"){'nation': 'United States', 'description': 'general', 'iso2': 'US', 'iso3': 'USA'}
You can also run it from the command line (multiple arguments allowed)
$ python -m flydenity AF1234 D-1234{'AF1234': {'nation': 'United States', 'description': 'general', 'iso2': 'US', 'iso3': 'USA'},'D-1234': {'nation': 'Germany', 'description': 'gliders', 'iso2': 'DE', 'iso3': 'DEU'}}
In total, the dataset contains a total of 408 unique regular expressions to
describe aircraft tail numbers across 217 unique countries.
Of course, everyone has a programming language of choice. Mine for this effort
was Python. I’ve including a wrapper class classed “ARP” which you can use to
parse through the expressions.
Since the ITU International Call Sign prefexies are universal across Aircraft
and Maritime Call Signs, we include functions within out API to parse Maritime
Call Signs as well.
To evaluate how well the regular expressions work, we extracted unique tail
numbers from a years worth of air traffic from FlightRadar24.com In total,
we evaluated over 250k unique tail numbers against the regular expressions
to minimize duplicate tags. In total, the parser was around 98% accurate in
matching tail numbers to a specific country. Of course this could be
improved, but that’s why this library is open-source :)
All data was collected using open sources across the web, specifically using
the links below.
I constructed two datasets (as of right now) for this effort.
Some of these countries or regions could have a standard that is not within
this database. Please update the list if you make changes.
Collen Roller
collen.roller@gmail dot com