Accidents are rarest in winter.
In https://www.ntsb.gov/_layouts/ntsb.aviation/index.aspx, click "Download all (text)", to get a file AviationData.txt
containing 80,000 accidents and 4,000 incidents.
Keep only the Accident
s, split the |
-delimited fields to get mm/dd/yyyy, then split by /
to get the month, then count how often each month appears. From a Linux shell:
grep Accident AviationData.txt | awk '{split($0,a,"|"); print a[4]}' | awk '{split($0,a,"/"); print a[1]}' | sort | uniq -c
Result:
4439 01
4697 02
5989 03
6632 04
7838 05
8758 06
9806 07
9223 08
7607 09
6267 10
4922 11
4485 12
Accidents are rarest in January, become smoothly more frequent until July, and then become smoothly rarer until December, i.e., the subsequent January.
But is this trend just because summer has more flights?
Let's compare apples to apples,
instead of 747s to everything including powered parachutes.
This table shows that airline passengers have the same trend,
about 26% to 31% more in July than January. Numbers that big argue that airline flights have the same distribution.
When filtering AviationData.txt to include only Boeing, Airbus, etc,
July is about 90% more than January.
So summer has 30% more flights than winter,
but 90% more accidents.
Even after adjusting for this, an airliner flight in winter is only a third as likely to have an accident as a flight in summer.