While working on wordpress/gutenberg#4072 today, I was inspired to do some data analysis on the WordPress.org plugin directory.
To prepare, I populated a plugins table with data from the WordPress.org REST API via this plugin-stats.php script. Download the SQL file to avoid needing to re-fetch the API data.
Based on the initial API request, there are 49,749 total WordPress.org plugins.
Of the entire data set, only 18,002 plugins have 200 or greater active installs. The remaining 31,747 plugins represent an inconsequential number of active installs compared to the total.
The 18,002 plugins represent 182,296,500 total active installs. A WordPress install can have multiple active plugins, so this total isn't unique WordPress installs. Also, we can ignore the remaining ~32k plugins because they would only represent 3,174,700 additional active installs if each plugin had 100 active installs.
Of the total active installs, 168,623,000 (92.5%) are represented by 3,440 plugins with >=5000 active installs. For that matter, 159,720,000 (87.6%) are represented by 2101 plugins with >=10000 active installs.
It'd be interesting to know what percentage of WordPress installs have a plugin not tracked in the WordPress.org plugin directory (e.g. premium or custom).
Interesting analysis. Sounds a bit like 10% of plugins represent 90% of activated plugins.
I wonder if those 90% seldom used plugins represent much benefit to the community?
This is very much to be determined. More so, popularity is used to determine where we should focus our efforts first.
Interesting numbers. The rule 80/20 applies well in this case. If we can categorize the plugins, we might know what users actually need.
I did something similar last year I thought you might interested in seeing: https://nabeel.us/2017/06/14/wordpress-org-analytics/
PS: Great job with the Changelog interview (that’s how I found this article).
Neat — and thanks!