I had some worries that aircraft might be counted in the FA stats as “new” if they would not have been received for some minutes and
would then reappear.
To investigate that i wrote a small python script which reads the data/aircraft.json of dump1090-mutability every 30 seconds.
Newly seen aircraft are counted by this script but only if they were not seen in the period of 15 minutes before.
To my surprise (and relief) it appears that the FA stats consistently report 10% less counts instead of more. So their filtering is even
more tight.
/paul
Aircraft that are only Mode S (lacking lat/lon information) are ignored when producing the ADS-B count. Aircraft are not considered to be “new” if they were previously seen in that UTC day. There is some other minor data validation to exclude extremely bad positions, but that’s mostly it.
Thanks for clearing that!
/paul
So what about aircraft with the same ICAO code but different flight numbers seen in the same day. For example, I often track an arrival to the gate where it sits for an hour and then departs with a new flight number. Would this be counted as two new aircraft or one?
The ICAO hex code is ignored for the purposes of counting aircraft in stats, which is only looking at the ident field. Aircraft that have the ident blank or unspecified will be ignored.