I’ve just updated:
and I see Python 3.11 loosing 8% of users in a month, while Python 3.9 is gaining 4%, this looks strange.
The overall total is also decreasing, which can be explained by the end of year holidays. it happens every years.
Interestingly, the decreasing overall total can explain, alone, the decrease in 3.11 usage.
So maybe professionals are using Python 3.11 and went on holidays, while hobbyists are more on Python 3.7 and 3.9 and went active during the holidays?
Perhaps a more accurate guess is that living individuals prefer to use 3.11 who go on vacation, while automated systems are still running 3.9 during the holiday.
The question you always need to ask is percentage of what?
Without the absolute download figures you have no idea what this means.
There’s an sqlite file in the repo you can query:
$ sqlite3 python-versions.sqlite
sqlite> select start_date, python_version, download_count from python_version where start_date >= '2024-11-01' and python_version in ("3.9", "3.11");
2024-11-01|3.9|8675239684
2024-11-01|3.11|13157223061
2024-12-01|3.9|10075275669
2024-12-01|3.11|8456510016
But if we want to be really sure we’d need to produce the same stats at a weekly or daily level, in this repo I only store monthly stats 