Every day, WordPress sites check for plugin updates. The API keeps a count of these update checks. From this, we determine “Active Installs”. This is basically a count of the sites that said "hey, I’m running Plugin X and want to check for an update for it. Then at the end of the day, we count those, and you have an Active Install count.
Simple so far.
Now, there is variation in that data from day to day. This is not the result of people installing or removing plugins alone; it also could be that a site just missed a day in checking for updates. Maybe a site was down for a day, or nobody looked at it that day so it never ran and never checked.
Thus the Active Install count is not an accurate count. It simply can’t be such. So we round it to the most significant digit and show that. This is because the data matters more for smaller plugins. You care when your plugin hits 80 installs, or 100 installs, or 200 installs, but by the time it has 10000 installs, then that 10020 just isn’t really a significant milestone for an author anymore.
Basically, instead of showing the exact data, we show something more useful, like trends. The growth chart is a chart of the change in the Active Install count over time.
For each week in the year (Sunday to Saturday) we say “get the average Active Install count for those 7 days”. Then for each of those values, we iterate through them to perform this calculation:
$growth = ( $this_week_average - $last_week_average ) / $last_week_average
This is rounded to the nearest 0.1 and shown as the percentage growth for each week.