For me, stars have an inverse correlation with actual usage. I'll star projects I want to look at some day, because they seem cool, like Svelte or Vite.

But projects I work with daily, I don't need to star them to remember they exits, like React or Webpack.

I am the maintainer of star-history.com. Thanks for posting this on HN.

Just to share a trick, you can also use markdown to embed a live star history chart to your GitHub README. e.g.

## Star History

[![Star History Chart](https://api.star-history.com/svg?repos=bytebase/star-history...

Instruction: https://star-history.com/#bytebase/star-history&Date

Enjoy~

Slightly related, but... what is it with stars? I've never starred a repo and don't understand why I would want to. Yet it seems to be a popularity contest people care about for some reason, even though it correlates strongly with forks/issues/prs so you get the same idea about usage from those. Does it go at all beyond a "like", or am I just too old to get it?

I use this a lot. I find it interesting to compare trajectories of similar projects. it's often hard to predict the popularity (or at least star count) of a library. Case in point, back when Bootstrap was in vogue I always found the Vue.js version of far superior to the react-strap (and react-bootstrap) in terms of API, documentation, and support. Yet bootstrap-vue never came that close to catching it's React counterparts in stars.

To play devil's advocate to those who thumb their noses at GH stars, I've found this metric to be a helpful proxy in my career when choosing libraries.

I get many stars on my last project but no issues or PR. It feels quite lonely and, more importantly, I can't know if my work is going in a good direction.

I think people mainly use stars as bookmark.

I wish Github were more permissive with traffic-related data to your Github pages. I'm not expecting GA tags, but being able to see > 14 days of page view history would be a good start.

Anyone know of any other solutions around this sort of thing?

Star can only show part of the truth, and a more comprehensive analysis may be needed to show popularity in multiple dimensions, like this website does: https://ossinsight.io/

Similar project that has ability to set time period to view, removes the hand drawn look, and better X/Y axis choice.

https://github.com/seladb/startrack-js

we are building a similar tool https://ossinsight.io/ which saves nearly 45B events from GitHub and can let you get more powerful insights from the project, e.g here is the history insights for K8s https://api.ossinsight.io/share/1bdadc17-6d94-41fb-8bc4-d6dc..., enjoy it by yourself :-)

I made a CLI version of a star history grapher for my own use, because I wasn't happy with the granularity of the graphs available in the various online incarnations of this.

https://github.com/dtolnay/star-history

Here is a side-by-side comparison of graphs generated by star-history.com vs my tool: https://github.com/dtolnay/star-history/issues/8

You can distinguish a lot finer structure in my graph: coinciding with individual blog posts with not as much reach as something on Hacker News front page. That structure is almost entirely concealed in star-history.com by the cartoony graphs.

If you want see the effect of HN on a new repository: https://star-history.com/#samwillis/tetra&Date

It was on the HN homepage for about 5 hours yesterday, never higher than about 25. The initial jump on the 25th was some attention on Twitter. Also the HN link was to the website, not the GitHub. I imagine if it had been a link to the repository the stars would be higher.

I wrote a tiny tool to calculate a "brightness" score for a repo, which looks at the total number of stars that the people who starred your repo have.

The idea being that people who starred your repo who themselves have tons of stars are likely "better" in that they are likely to be experienced, intelligent users, and not bots.

https://github.com/Hellisotherpeople/Bright

I have a project that allows not only to compare the stars, but also to find similar repositories, trends and interesting facts: https://ghe.clickhouse.tech/

Very cool. I had no idea GitHub's API kept history statistics for stars!

Tutorial on building it with Python https://youtu.be/TzF-OUA1Tlo

https://www.npmtrends.com/ if you want the same for npm data.

Github has always had a bad history graph. I've always wondered why we can't just have a live Github history graph like in Sourcetree.

Anyone know how they're creating such a cool chart like that?

Love the xkcd style chart!

<shameless-plug> We also show the GitHub star history, alongside npm download history, issue and PR stats on Openbase: https://openbase.com/js/react/insights </shameless-plug>

this is really cool. nice work

i use this thing all the time, but wish i had the option to turn off the XKCD style charts to some thing more "professional"

Is that a pentagram?

It makes me feel good when I see one of my projects getting stars. When I star a project it is with the hope that another maintainer will have a similar good feeling.

I think of it as a "like" or a bookmark, if I come across a repo and I find it interesting or useful, I star it.

GitHub recently added a feature where you can organize your stars in lists, so you decide what they mean. My lists include for example "useful CLI tools", "languages", and "projects to contribute to". Outside of this feature, I mostly use the stars as a 'like' button.

It's a 2-in-1 like & bookmark.

I wrote a blog about how stars helped us at the very early stages of our open source product: https://www.infracost.io/blog/github-stars-matter-here-is-wh...

> Yet it seems to be a popularity contest people care about for some reason

Before you incorporate someone's Github library into your own project, you want to know some things. Is it secure? Is it well-maintained? Is there an active userbase to help with support? Stars are one indicator of this.

I use them mostly like bookmarks, a cool project I stumbled over and want to "pin" for the future.

I use it as a bookmark. Under my own account I can find the repos I have starred.

I recommend adding telemetry to your project for this (and I know a lot of people feel strongly about this, so I'll add: with a very easy way of disabling it).

In OctoSQL[0] I'm literally just sending JSON files with coarse information about 1. invocations of the CLI, 2. features used in these invocations, to a VM on DigitalOcean (with a 10-line server receiving them and writing to a JSON file - which I can then process using OctoSQL itself).

Thanks to this I knew that until recently most of what I had were stars, not actual usage, and could also see how the big rewrite I released in January (and the following updates) made a lot of people start actually using it since. Very nice feeling :)

All telemetry logic in OctoSQL is actually contained in a single short file[1].

PS: Keep in mind that https requests take a long time, as they need to do a few roundtrips for the TLS handshake. Don't do that on every invocation if you have latency-sensitive invocations. I.e. in OctoSQL I'm only sending aggregated telemetry data every 10 invocations (as well as on the very first one).

[0]:https://github.com/cube2222/octosql

[1]:https://github.com/cube2222/octosql/blob/main/telemetry/tele...

I think of stars as encouragement, so I give them when it has fewer than a few-k stars and is well maintained, or if not, to encourage it.

Is there an established way that people advertise their open source projects to developers for help? I know that GitHub has exploration features but when I last used them it felt like looking for a needle in a haystack.