What problem does it solve? Designing a Neo4j graph schema without established patterns leads to redundant data, generic relationships, missing constraints, and slow queries. This Skill provides proven modeling conventions and a reference fraud-detection schema so you can build correct, performant graph models from the start. ## Core Features & Use Cases - Modeling Conventions: Enforces CapitalCase labels, UPPER_SNAKE_CASE relationships, camelCase properties, and anti-pattern avoidance (symmetric relationships, foreign keys as properties, label overload). - Reference Fraud Detection Schema: Provides a complete banking/fraud investigation model covering Customer, Account, Transaction, Device, Session, IP, Alert, and Case nodes with constraints, indexes, and vector search for facial embeddings. - Data Loading & Performance Guidance: Covers batch loading with MERGE, unique constraints on business keys, index strategy, and query profiling. - Use Case: When building a fraud detection system, apply the canonical transaction schema to model customers sharing emails or devices, then traverse the graph to detect linked identities and open investigation cases. ## Quick Start Use the neo4j-data-models skill to design a graph schema for my fraud detection application with customers, accounts, and transactions.
