Python 3 now has 75% adoption and Python 2 has 25%, at least according to the Python developers survey 2017 referenced on Python.org, however google trends doesn't seem to show this.
The following graph is from google trends, comparing 2.7 vs 3.4, 3.5, 3.6, and 3.7. I would expect to see a much greater difference if 2.7 has only 25% usage.
It does appear in the end of the 12 month graph that Python 3.6 is finally about to surpass 2.7, but has not yet done so. The 5 year trend shows that both 2.7 and 3.6 are both growing but 3.6 is catching up quickly.
It's interesting that the data in the latest report does not seem to correlate to the 5 year graph. While "in 2016 60% were using Python 2 compared with 40% for python 3" is clearly shown in the graph, the latest report of 75% using python 3 vs 25% using python 2 does not appear to be accurate. It appears to be more of a 50% - 50% or 40% - 60% split if you account for all the python 3 versions.
I use duck duck go, and Python 2.7 results still always show up first. What do you think? Did 3.x finally actually pass 2.7? What explains the disparity between the report and google trends data?
If you're wondering, I use both versions.
The following graph is from google trends, comparing 2.7 vs 3.4, 3.5, 3.6, and 3.7. I would expect to see a much greater difference if 2.7 has only 25% usage.
It does appear in the end of the 12 month graph that Python 3.6 is finally about to surpass 2.7, but has not yet done so. The 5 year trend shows that both 2.7 and 3.6 are both growing but 3.6 is catching up quickly.
It's interesting that the data in the latest report does not seem to correlate to the 5 year graph. While "in 2016 60% were using Python 2 compared with 40% for python 3" is clearly shown in the graph, the latest report of 75% using python 3 vs 25% using python 2 does not appear to be accurate. It appears to be more of a 50% - 50% or 40% - 60% split if you account for all the python 3 versions.
I use duck duck go, and Python 2.7 results still always show up first. What do you think? Did 3.x finally actually pass 2.7? What explains the disparity between the report and google trends data?
If you're wondering, I use both versions.
My guess is that although people are preferring Python 3 for new development, the majority of *running* software is still on Python 2. The PyPI download statistics should be a good indication of the latter: you can query them at https://bigquery.cloud.google.com/dataset/the-psf:pypi?pli=1 :
ReplyDeleteSELECT
substr(details.implementation.version, 0, 3) as ver,
COUNT(*) as downloads,
FROM
TABLE_DATE_RANGE(
[the-psf:pypi.downloads],
TIMESTAMP("20180401"),
TIMESTAMP("20180407"))
GROUP BY ver order by downloads desc
1 2.7 96835255
2 3.6 19353614
3 null 18605539
4 3.5 10602418
5 3.4 4737409
6 2.6 940158
7 2.4 146161
8 3.7 107378
Compared to the same period in 2017:
1 2.7 100867390
2 null 35321706
3 3.5 9847261
4 3.4 5853988
5 3.6 4433603
6 2.6 3319993
7 3.3 294804
8 2.4 106793