"Who's the strongest wrestler?" can't be answered by win counts alone. Chaining graph algorithms to surface true dominance is a fun read 🥋 Title: SumoDB in Neo4j: Chaining Multiple Graph Algorithms in Snowflake — Part 3 URL: 🥋 Overview This post combines Neo4j Graph Analytics with Snowflake SQL to measure "dominance you can't see from win counts" in professional sumo data. It chains multiple graph algorithms into a composite "Chaos Score." ❓ Challenges Solved Ranking by raw wins overrates wrestlers who just beat weak opponents. By pairing Neo4j and Snowflake, the post surfaces competitive structure that neither tool alone could reveal. 💡 Methodology & Proposed Approach It builds weighted directed "winner → loser" edges and chains three algorithms. ・PageRank: weights wins over stronger opponents higher, measuring victory quality ・Betweenness centrality: finds bridge wrestlers connecting elite and mid-tier ・3-cycle detection: visualizes rock-paper-scissors (non-transitive) rivalries A damping factor of 0.85 and reversed edge orientation direct prestige toward winners, converging in ~20 iterations. 🌍 Use Cases ・Talent assessment: separate inflated win records from genuine dominance ・Structural analysis: find key wrestlers whose removal fragments the hierarchy ・Competitive balance: gauge ecosystem health via non-transitive rivalry density #GraphDataScience# #Neo4j#