17

Put simply, is there any way to find out how widely used a given open source project is?

Say, for example, I am looking for a serialization library. I can imagine various metrics to help choose among the many available, but a major one would be, "How widely used is this?"

Wide adoption tends to suggest either a very mature project, or one with a very active community.

For Python, PyPI at least provides download statistics. On GitHub you can see the number of forks as a guideline.

For a developer embarking on a new project, and wanting to leverage an OSP, this type of information is very valuable in making their choice.

The open source movement is not just about publishers, but consumers as well.

5
  • 4
    I'm voting to close this question as off-topic because it particularly asks for an outside resource providing statistics on the usage of an open source project, and is not necessarily directly related to open source itself.
    – Zizouz212
    Commented Jun 23, 2015 at 19:22
  • A good indicator are popcontest of Debian and Ubuntu at e.g. popcon.debian.org
    – frlan
    Commented Jun 23, 2015 at 19:24
  • 1
    I agree it is partially off topic, but I think it is one that many people coming to an SE site about OSS are likely to be wanting an answer to. Perhaps rather than close it, it can be suitably edited. Take away the seeming ask for a 'recommendation' for example.
    – kdopen
    Commented Jun 23, 2015 at 19:37
  • 1
    statistics for an open source project is not necessarily directly related to open source......wat
    – albert
    Commented Jun 23, 2015 at 22:58
  • @kdopen Perhaps, but then we should add a notice to it so that we don't get duplicates of off-topics. Of course, that will likely come when we go public, since everyone here is aware of the community.
    – Zizouz212
    Commented Jun 24, 2015 at 1:59

8 Answers 8

8

There's no way of actually tracking down usage statistics without compromising the privacy of its users. Hence why it is impossible to properly see which Linux distribution is the most popular one (using this as an example).

You could use some analytics like the number of downloads, the number of stars on GitHub, or the number of contributors to the project, but those analytics can in no way properly determine how broadly is the open source project actually used.

But it might be possible to do so for the corporate environment in some cases. Open source projects could theoretically keep a list of organizations using their product. Also, if the project is mentioned multiple times in the media, there's a higher chance that there are many users using it.

6

It depends what you mean by usage statistics, but the website www.openhub.net (formerly known as Ohloh) gives you:

  • the number of commit per months
  • the number of contributors per month
  • a community rating
3
  • 1
    Unfortunately, OpenHub became slower and slower over the last months. Right now it has reached a point where it is practically unusable.
    – vog
    Commented Aug 6, 2015 at 21:54
  • I've put this site in a proper answer but to reply to @vog - a newer site with a similar aim is stackshare.io.
    – RustyFluff
    Commented Jan 11, 2019 at 17:19
  • @RustyFluff Thanks! I didn't know about Stackshare, sounds promising.
    – vog
    Commented Jan 12, 2019 at 20:20
2

I haven't come across any single, centralized place where all usage statistics for open-source projects are combined.

However, all is not lost: many software download sites have a download counter. Instead of looking centrally, try going to the software's download page and looking for any stats it might have there.

Similarly, I haven't seen any libraries that do the job for you, but if you were to write your own (and release it as open-source!) then you could take the download location as an input and scan through it for statistics to return.

2

Every platform has its own metrics and indications.

npm for example, shows download statistics. GitHub has stars, which is a good rough indication.

The fact is, I don't need to host the file on some fancy VCS host for it to be considered open. I can host it on a server I control, and distribute my git URL for you and the public to clone and use. In which case, you will never have any sort of statistics unless I specifically make some.

1

I understand that SF isn't popular anymore, but they are sitting on a valuable pile of historic data. For example: http://sourceforge.net/top

1

Repology can be a useful way to show in which software packaging repositories a given project is packaged for, which indicates that users have created packaging control file(s) for the software, a non-trivial task, and have it accepted by the maintainers of the packaging system, again not trivial. Then, how often are those control files being maintained can give a unique perspective on how widely used the software is.

For each packaging system which is distributing the software, it is useful to determine whether it hosts all of the files, in which case the upstream download links will not see those users. Packaging system might provide some stats for downloads by its users, but many do not.

e.g. https://github.com/nuitka/nuitka is https://repology.org/project/python:nuitka , which shows a handful of distrubutions have built packages for it, and they are all not keeping their package updated.

Compare with a more common package like pyflakes: https://repology.org/project/python:pyflakes/versions , appearing in lots of packaging distributions, and many are up-to-date.

It includes many of the C/C++ packaging systems (e.g. Vcpkg), and Windows packaging systems (e.g. Chocolatey), so isnt only Linux Distros. e.g. https://repology.org/project/azure-cli/versions and https://repology.org/project/7zip/versions

0

You can build a website that hosts your project's online documentation, latest news, etc. You can then track the relative popularity over time, based on traffic. You can use this to indicate increasing popularity, or trends for certain OSes or browsers - typical web traffic analysis.

But regardless of any specific download-tracking mechanism, you are treading on your user's privacy, and that's not likely to be well received in the open source community.

0

Old question but I have a new genuinely useful site that has grown to become very pertinent here.where

stackshare.io

Its a site which gathers data from Github and generates scores based on social media activity. It also has user accounts that enable you to list what software stacks you use in your day job/business. Thereby adding to the data.

Very good site similar to openhub.net

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.