\

Show HN: GetStack.dev โ€“ Track GitHub open-source trends

42 points - last Friday at 7:50 AM


Hi HN!

Iโ€™ve been working on getstack.dev[1], a tool to help developers track GitHub open-source trends, tech adoption, and repository stacks โ€” updated weekly.

About a month ago, I broke my leg. While stuck on the couch, I figured Iโ€™d put the downtime to good use and finally build a side project Iโ€™d been thinking about for a while. So I put together an MVP and decided to release it publicly to gather feedback.

I have always struggle to grasp how people are adopting technology and what's really hype or under the radar. As tech leader you also often want to know if your tech choices are the right one but it's hard to take a data driven solution. And as open-source lover I always want to know how my favorite projects are built.

All the data is pulled and refreshed weekly from GitHub, stored on ClickHouse [2] but you can directly check how I built it in the website [3]

[1] https://getstack.dev

[2] https://getstack.dev/tech/clickhouse

[3] https://getstack.dev/specfy/getstack

Source
  • mlhpdx

    yesterday at 3:05 AM

    Iโ€™m not sure how to take โ€œopen sourceโ€ when there are closed source commercial things on the trends list.

      • h1fra

        yesterday at 8:42 AM

        It's true, but source available is less sexy and understandable

    • herpdyderp

      last Friday at 11:10 PM

      Some of the categories don't make sense to me. Angular is not a language, neither is Deno, for example.

        • h1fra

          yesterday at 8:48 AM

          I agree, sometimes it's hard to put a label that would fit what most people expect. In the case of Angular I have put it in languages because I put React there too, React is there because it has a specific file extension and language. I could put them in framework but they would be mix with a lot of stuff that people do consider framework and not languages.

          Same for Deno (and nodejs and bun) they would fit better in a Runtime category maybe but I'm not sure people would understand that category and that it would make a meaningful comparison.

          • chrisweekly

            yesterday at 3:55 AM

            I sympathize w OP; the ecosystem doesn't always fit into a clear ontology. That said, you're 100% right that those 2 examples were miscategorized.

        • exiguus

          yesterday at 11:34 AM

          What is the value for me as a software engineer to watch this?

          E.g. How does the metric work.

          • Brajeshwar

            yesterday at 4:15 AM

            You are hot on Hacker News, but your newsletter subscription is failing with, "An error occurred"

              • h1fra

                yesterday at 8:41 AM

                Thanks there was an issue with the Api key

                  • Brajeshwar

                    yesterday at 8:56 AM

                    Cool. Subscribed.

            • lawgimenez

              yesterday at 2:05 AM

              I searched for Kotlin repository but it was not found. It has like 50K stars.

                • h1fra

                  yesterday at 8:52 AM

                  The repo has been ignored from compilation because it's too big to parse in a reasonable time, all of this is quite costly. Tried to be smart by filtering the search but maybe it's confusing in that case

                  https://getstack.dev/JetBrains/kotlin

              • Sourabhsss1

                last Friday at 8:25 AM

                This is good stuff.

                • xp84

                  last Friday at 11:39 PM

                  I feel like the way you present "trends" like on here: https://getstack.dev/category/language

                  ...is less than useful, because (roughly) no one deletes old repos and code, so everything will always be trending "up."

                  I'd be more interested in a stat that perhaps considered "number of active repos that have this language, or, this language's share of representation among repos with activeness in the last month. With some reasonable definition of "active," of course.

                    • h1fra

                      yesterday at 8:44 AM

                      It's more or less that, the about section answers this. It's parsing popular repo that has been active in the last 2 years. It's a long period but otherwise the trend would move too much that would make them useless and a lot of popular repo are not maintained or don't need regular update

                  • herpdyderp

                    last Friday at 11:08 PM

                    The percents are... change per week?