Quine

Hello ProductHunt 👋 Today we’re excited to announce that we’re launching QuineTopics, a taxonomy of open source derived automatically from READMEs and designed to make the ecosystem more organised and accessible! At Quine, we're quantifying experience and reputation of software developers from code and open source metadata. A central question is knowing what areas a developer is an expert on. This is a difficult problem given that we don't even know what "areas" exist in open source. To solve this problem we created Quine Topics, which uses NLP to automatically infer the labels of a GitHub repo from its README and Description metadata. Through this approach we have reduced the label space to a controlled vocabulary of ~600 (a decrease of 400x), and automatically tagged software projects via use of machine learning. Our taxonomy has a 700% decrease in unlabelled projects versus GitHub. Additionally, we created a 2D map of the space to help developers explore and visualise the repoverse by allowing them to browse the top projects and contributors in each category. You can learn more about our approach at https://medium.com/@alxbd/a-taxo.... Quine Topics is a milestone in our journey to solve the attribution and reputation problems in open source, and will play a fundamental role in many of the products we build next. We’re very interested in your feedback! If you like what we're doing, come hang out with us at https://discord.gg/quine.

Good one.

Taking ~240,000 "GitHub topics" down to ~600 "Quine topics" will make navigating the open source ecosystem and finding projects and developers much easier! Looking forward to what this foundational work can unlock in the future. Exciting times ahead 🎉