@Schlueter: There is Stack Overflow, Computer Science, Theoretical Computer Science, Code Review, Software Engineering, Data Science, Cross Validated, Computational Science which all somewhat fit. Especially if one poses this as the clustering problem that it really is, the data science and machine learning sites becomes even more relevant. If one would find an embedding per activity, then one could use cosine similarity and appropriate data structures. At this point of Stack Exchange, I'm honestly at a loss when trying to find the correct site to ask. —
Martin Ueding 57 secs ago