By Zoran Pandovski
Python package download stats from PyPI
The download stats were removed from PyPi modules. There were few reasons for that as explained in the mailing list. So, now the best available option to get the statistics is to use the Google BigQuery database. To use it you only need a Google account and enabled BigQuery API.
The most useful query is to count the package downloads:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
SELECT
COUNT(*) AS num_downloads,
SUBSTR(_TABLE_SUFFIX, 1, 6) AS `month`
FROM `the-psf.pypi.downloads*`
WHERE
file.project = 'YOUR_PACKAGE_NAME'
AND
details.installer.name = 'pip'
AND _TABLE_SUFFIX
BETWEEN FORMAT_DATE(
'%Y%m01', DATE_SUB(CURRENT_DATE(), INTERVAL 10 MONTH))
AND FORMAT_DATE('%Y%m%d', CURRENT_DATE())
GROUP BY `month`
ORDER BY `month` DESC
The above query uses the the-psf.pypi.downloads table that contains information about downloads. It selects the package and filters the data to contain only pip installs(no mirrors downloads) in the last 10 months.
Other useful columns from the the-psf.pypi.downloads are:
| Column | Info | Example |
|—|—|—|
| file.project | The package name | requests |
| file.version | Version of the package | 1.2.1 |
| details.installer.name | Installer name | pip, bandersnatch |
| details.python | Python version | 2.7, 3.6, 3.7.3 |