Citizen observatory of aircraft overflights
How the figures are measured
Every figure on this observatory comes from a written method. Here it is in full, including what it measures and what it deliberately does not.
What counts as an overflight?
An aircraft passing above the outline of a zone. The outline is not the municipal boundary: it is the strip that lies under the flight path, defined and verified before the zone exists on the site. That is why some zones are not here yet, and why none appears with an approximate coverage in the meantime.
A single flight crossing two zones counts once in each. They are two different exposures, lived by different people, and adding them into one total would state something nobody experienced. Whenever you read a figure, check which zone it belongs to.
Why is the altitude shown the lowest one?
Because it is the one you feel. Of all the positions recorded while the aircraft was over the zone, the lowest is published: it is the real ceiling above that house, not an average of the track.
Read it as a ceiling, not as an average. An aircraft that comes in high and leaves low did not pass at middling height: it passed low.
How often is it updated, and how much history is kept?
Positions are collected every hour, all day. Noise readings do not arrive at the same time as the pass: they are published later and their classification settles over a few hours, so an hour is only treated as closed once it has been checked to have stopped changing. Nothing is held back on purpose: what is there, is there.
This observatory’s archive deletes nothing. Every recorded pass and every noise reading is kept permanently, and the history grows day by day from the moment the zone opened. The only thing emptied is the antechamber where a pass waits a few hours to be confirmed, and only once the permanent record already holds it.
What is limited is how far back you can look before that date: at the start only the preceding 60 days could be recovered, because beyond that the data was no longer available anywhere. From then on nothing is lost.
Why are there thirty passes some days and a thousand on others?
Because the runway configuration changes with the wind, and with it changes where departures and arrivals go. This is not small variation around an average: they are two distinct regimes. The same zone can see tens of passes one day and on the order of a thousand the next.
That makes a “daily average” a misleading figure, which is why none is published. Measuring a single day says little on its own: it tells you which regime that day was in. What does mean something is the distribution across the window.
Why is there noise recorded with no aircraft nearby?
Because noise does not stop at the edge of the zone. Fixed monitoring stations record what reaches them, and it reaches considerably further than the aircraft passes: there are days with dozens of noise readings and only a handful of passes over the outline.
Those readings are not deleted or hidden. They are part of what is measured, not a flaw in the method: publishing only the noise that matches a pass would erase half the phenomenon from the record.
Can everything recorded be used to file a complaint?
These are two separate things and they are kept separate. Everything recorded counts towards the statistics, always, unfiltered. Whether a particular episode also meets the conditions for filing a complaint is a different question, and it changes none of the published figures.
When an episode meets the conditions, the site does the hard part: it prepares the draft with the date, the time, the zone, the aircraft model and the lowest measured altitude, and adds the wind paragraph when that hour’s conditions support it. Nobody has to write that at eleven at night.
The complaint is still always personal: you file it and it goes in your name. The draft can be copied and pasted, or sent from the site itself with a copy to your own inbox. This observatory does not file on anyone’s behalf.
What does this observatory claim, and what does it not?
It does check whether an episode may be actionable. Every pass is cross-checked against the published rule —the band it sets and the written conditions that allow departing from it— and against that hour’s weather report. When none of those conditions is on record, the draft says so, and says it as a question: not “they flew badly”, but “that hour’s report listed none of the stated causes; which one was applied?”.
What it does not do is declare an infringement. Publishing how many passes there were, at what height and with how much noise describes an exposure; ruling on it belongs to whoever has the authority, which is why what is explained is how the rule is designed and never how anyone behaved.
Nor does it promise the record is complete. A pass can be missing and a reading can arrive late. What does hold is that what appears, did appear: no figure is adjusted, conveniently rounded or delayed to look better.