diff --git a/content/download-stats.html b/content/download-stats.html new file mode 100644 index 00000000..9732b600 --- /dev/null +++ b/content/download-stats.html @@ -0,0 +1,525 @@ +--- +title: JBang Download Stats +tagline: GitHub release download statistics for jbangdev/jbang +link: /usage/downloads/ +layout: splash +--- + +{| +
+ Live download statistics for + jbangdev/jbang + GitHub releases, fetched straight from the + GitHub Releases API + in your browser (inspired by grev). +
+ +
+ Two things are done differently here:
+ version.txt is separated out — it is the tiny file
+ JBang downloads on every
+ update check,
+ so its count dwarfs real installs and would otherwise skew the picture — and the
+ x-axis follows real time (each release sits at its actual publish date),
+ not evenly spaced release tags, so bursts and quiet periods show honestly.
+
+ Total downloads each release's binary artifacts have accumulated to date, placed at the
+ moment the release was published. Signatures (.asc), checksums and
+ version.txt are excluded.
+
+ Downloads of the version.txt asset — overwhelmingly automated update
+ checks, not installs. Shown on its own axis so it never distorts the installer chart above.
+
| Release | +Published | +Installers | +version.txt | +Checksums / sigs | +Total | +
|---|
+ Note: GitHub reports each asset's cumulative download count to date, so a bar shows + a release's lifetime downloads positioned at its publish date — not activity within that + period. Data is cached in your browser for a few hours to stay within the GitHub API rate limit. +
+This is a work in progress. The data is collected from the JBang version checking here.
+See also the GitHub release download stats (installers vs. version.txt update-checks, over time).
The data is updated once a day. It will show data for maximum last 3 months, starting from 2024-10-19.
% is calculated as unique per approximate location info used for the http request. Meaning if the same approximate area makes multiple requests using the same version/country it is counted only once.
NOTE: The data is an approximation. This is not seeing all invocations, just those that reach the version check endpoint. Meaning nothing behind proxies, nothing that runs jbang very fast (as in no time to do an update check) or just don't do an update check. The total count is unknowingly higher - thus just use this as a interesting data point - mainly for fun!
diff --git a/content/usage.html b/content/usage.html index 6a8554fe..19556f9c 100644 --- a/content/usage.html +++ b/content/usage.html @@ -21,7 +21,8 @@You can find some other fun stats in the Leaderboards.
+You can find some other fun stats in the Leaderboards + and GitHub download stats.