A field guide to the sprawling science of graph clustering

What it does

This repository is a curated index of academic papers on community detection, organized by methodology and paired with implementations where available. It spans thirteen distinct algorithmic families — from spectral methods to hypergraphs — serving as a shared reading list and code reference for the field. Think of it as a card catalog for graph clustering research.

The interesting bit

The value is entirely in the taxonomy. The maintainer splits a sprawling literature into precise chapters like “Cyclic Patterns” and “Centrality and Cuts,” which is more navigable than a flat dump of arXiv links. The companion lists for graph classification and fraud detection suggest someone is trying to map the whole network-mining territory.

Key highlights

• Thirteen specialized chapters covering distinct methodological families (factorization, spectral, temporal, block models, etc.)
• Explicit focus on papers that include implementations, not just theory
• CC0 license, so the index itself is public domain
• Companion “awesome” lists for related topics (graph classification, fraud detection, gradient boosting)
• 2,400+ stars suggest the graph-mining crowd finds it genuinely useful

Caveats

• The main README is just a skeleton; the actual paper lists live in separate markdown files, so browsing requires clicking through chapters.
• No inclusion criteria, last-updated date, or maintenance cadence is visible, so freshness and vetting standards are unclear.

Verdict

Worth bookmarking if you are doing a literature review or choosing an algorithm for a graph-clustering problem. Skip it if you are looking for a single reusable library — this is a map, not a toolbox.