How to interpret github clone stats

Question

On github you can view the traffic to your repository like so:

Now my question is about the huge discrepancy in clones and unique cloners. If I understand this correctly, it would mean that 4919 people have created 22.374 clones. Why would some people clone a repository so many times?

My actual repo looks a lot less impressive, but the discrepancy is still big. (About 30 unique clones for 150ish clones, so each cloner takes 5 clones).

That makes me wonder what is counting as a clone. If it's relevant, it's for a go library so I am assuming a go get counts as a clone, and go get -u possible for another one each time they update. But it'd surprise me if people were updating this much in a 2 week window.

Answer 1

I think the most likely explanation is the following:

both git-pull and git-clone actually call git-fetch under the hood, in addition to doing the work relevant to their commands.

see first couple paragraphs of each's manpages for explanations.

i doubt github tries to differentiate fetches that are part of a git-clone from ones that are simply appending an existing repo. (i'm not even sure that would be possible, haven't given it much thought)

From what I've read there is not much information on this floating around, so this is a really good question. And as @Bobulous says in his comment if it's not explained in the official documentation and blog posts you are most likely not going to find an objective answer on this question here. All we can do is speculate on what is going on underneath Github's hood.

In any case I recommend reading the following links that might help you find your answer:

• The github stats on "clones":"unique cloners" seems heavily skewed, is "clone" actually "pull"?
• Explain clones info in Github's Traffic tab