Is their a way to automate reporting on GitHub stars & contributions?

#1

it would be cool to quantitatively report on open source contributions. one way to achieve that is to rank data from GitHub over time (like contributions and stars).

#2

Are you thinking about this for showing a stat on contributor bios? I like the idea. “contributions” could mean a lot of things though. Is this the unique number of repos someone has contributed to? It could also mean the number of commits, likes or something else.

So what if we were to define stars like this…contributors get stars based on their share of commits. So say a repo has 100 stars and 100 commits. If one contributor has 50 commits they’d have 50 stars and if another contributor has 10 commits, they’d have 10 stars.

I’d also like to see a commit streak in bios. A leaderboard of HN contributor commit streaks would also be nice.

#3

I was thinking more from the perspective of data storytelling for individual posts, but on the bio could also make sense. Also especially like the idea of celebrating a streak of consecutive open source contributions.

Back to the data, it could be a filterable leader board, or it could be raw data that contributors have access too. Posts like “Top 100 Open Source Contributors This Week” or “Top 100 Open Source Contributors in 2019” or more in depth reporting like “Open Source Contributions to Python Machine Learning Projects Are Up by 45%” or even predicting the the future adoption of the software/framework/language/project by measuring/comparing the volume of ppl contributing.

#4

I would suggest we define some equivalence with Github as follow:

  1. Each HN post is equivalent to one Github “repository”
  2. The readers can make “pull request / commit to a post” a.k.a request edit (it can be typo or references, …)

As far as I know, Github counts one “contribution” to be one of the following case:

  1. Create a repository === Published a new post. We can raise the activity point when it is Peer-reviewed as I suggested in this (What are the killer features you'd like to see in Hacker Noon 2.0?)
  2. Make a commit to a repository ===HN equivalence=== Accepted edit to a post
  3. Raise an issue. This one I haven’t figured out a way to map it into HN but it can be a highly favorited comment to a post?

We can also make a heatmap to visualize the contribution to HN of one user.

5 Likes
#5

Hey @Dane, @hackernoondev, and @austin, check this out https://git-stars.netlify.com/ it displays trending projects by stars in CSS, Javascript, Ruby, Python, Java and PHP. Source code: https://github.com/fayazara/git-stars

There is also opportunities to search repositories this way too: https://developer.github.com/v3/search/#search-repositories

I think reaching a simple and systematic way to rank GitHub trending projects would be good media for the software development community.

1 Like
#6

Definitely agreed! I would absolutely love to experiment with the Github API post-launch.

I had a project that used it several years ago, called “git-calendar”, that would give you an “x” on the calendar for every day you pushed a commit to Github (this was before they had the contribution graph on the profile page). Some sort of integration between Github and npm would be pretty nifty for selecting an appropriate open source tool for a given task…I’m picturing some metric that factors in npm downloads, github stars, and recent contributions to give you a score (like npm tries to do).

@Dane any thoughts there?

1 Like
#7

I’m game. I started a side project called Shifty a few years ago designed to let people subscribe to curated release notes for open source projects. I didn’t pursue it because I needed to build a community to do the curation before it would be valuable. Maybe this could be a nice labs product for Hacker Noon.

2 Likes
#8

Short, intelligent writeups about this page https://github.com/trending could be also be beneficial

1 Like
#9

This could be achieved by utilizing the GitHub API https://developer.github.com/v3/repos/ and ingesting the data into Elastic Stack or Elasticsearch Service on AWS. The integration with Kibana provides a vast array of data visuals.

#10

You might find this site interesting/useful for that purpose.
https://git.news/