2021-10-03 13:42:26 +00:00
|
|
|
# Analysis of version adoptions on PyPI
|
2021-10-03 13:39:45 +00:00
|
|
|
|
2023-06-15 13:30:46 +00:00
|
|
|
## pypi.org downloads by version
|
|
|
|
|
|
|
|
![](python-versions-lines.png)
|
|
|
|
![](python-versions-stacked.png)
|
|
|
|
|
|
|
|
|
|
|
|
## pypi.org downloads total
|
|
|
|
|
|
|
|
![](pypi-download-counts.png)
|
|
|
|
|
|
|
|
|
|
|
|
## Where does the data come from?
|
|
|
|
|
2021-12-29 10:24:21 +00:00
|
|
|
We get
|
|
|
|
[publicly available PyPI download statistics](https://bigquery.cloud.google.com/table/bigquery-public-data:pypi.downloads)
|
|
|
|
on Google BigQuery using [pypinfo](https://github.com/ofek/pypinfo/).
|
|
|
|
|
|
|
|
[Here](https://github.com/pypa/linehaul-cloud-function) is the repo
|
|
|
|
for the open-source code pushing the pypi stats to BigQuery.
|
2021-10-03 13:39:45 +00:00
|
|
|
|
|
|
|
|
|
|
|
## Usage
|
|
|
|
|
|
|
|
First you need to get an access to PyPI's BigQuery, by following
|
|
|
|
[pypinfo](https://github.com/ofek/pypinfo/) procedure.
|
|
|
|
|
|
|
|
Then there's two main invocations, first fetch the data using:
|
|
|
|
|
|
|
|
python python-versions.py --fetch
|
|
|
|
|
|
|
|
Then plot it using:
|
|
|
|
|
|
|
|
python python-versions.py
|