For better understanding: It is impossible to get correct Git code repository (not: Gerrit) statistics as we import a bunch of upstream repositories (and their git history) but then apply custom changes only for Wikimedia.
Example: Local merged patches in https://gerrit.wikimedia.org/r/#/q/project:integration/jenkins-job-builder vs the entire history of both local and upstream merged patches in https://phabricator.wikimedia.org/diffusion/CIJJ/history/master/

The Git (not Gerrit which is for code review, and in Gerrit a proposed code change can get accepted, abandoned, or remain unreviewed) stats at https://wikimedia.biterg.io/goto/ae37bc7715bea198d0301fd9c3b4dafb (excluding a bunch of repositories again which is not a definite perfect list either and might be incomplete as our software stack changes) show 50904 commits and 420 authors. That link uses a list of excluded repositories which we also applied in the previous two audit report editions and which I don't trust as per reasons outlined in the first paragraph here.

Which made me wonder whether to use the numbers for merged code changes in Gerrit instead of Git stats. (Note to myself: must set author_bot:false to exclude bots, e.g.L10n-bot which automatically merges updated translations.)
That's https://wikimedia.biterg.io/goto/b6b44987ee5deec0becd728a8a655870 which says that 48652 commits by 413 human authors got merged into Wikimedia Git repositories between 2017-07-01 and 2018-06-30. Those are the numbers I trust enough to go with.
Note that these numbers are not only volunteers but anybody. If numbers for volunteers-only were wanted, you could click the part called "Independent" on the pie chart.
Note that some code repositories do not use Wikimedia Gerrit but third-party external websites such as Github which we currently do not track due to T109939 blocking T186736.