Use Case Guides

Real-world domain modeling patterns and implementation guides for common graph database applications.

Available Use Cases

Fraud and Anomaly Detection

Detect fraudulent transactions and anomalous behavior using graph patterns, ML embeddings, and real-time analytics.

Key features used:

  • Graph algorithms (centrality, community detection)
  • Real-time analytics with CDC
  • Row-Level Security (RLS) for multi-tenant isolation
  • Audit logging for compliance

Suitable for: Financial services, e-commerce, payment processing

Recommendation Systems

Build personalized recommendation engines using collaborative filtering, embeddings, and vector similarity search.

Key features used:

  • ML graph embeddings (Node2Vec, GraphSAGE)
  • HNSW vector similarity search
  • Graph traversal patterns
  • Real-time updates

Suitable for: E-commerce, content platforms, social networks

Knowledge Graphs

(Coming soon) - Organize and query structured knowledge with semantic relationships, inference, and entity resolution.

Key features used:

  • GQL pattern matching
  • Full-text search with BM25
  • Schema constraints
  • Distributed query coordination

Suitable for: Enterprise search, data cataloging, research platforms

Common Patterns

Multi-Tenant Isolation

Use Row-Level Security (RLS) to isolate data by tenant:

-- Create RLS policy for tenant isolation
-- Not expressible as a Geode policy predicate (only `=` against a string
-- literal compiles); precompute the decision into a property instead:
CALL geode.rls.add_attribute_policy('app', 'Label', 'access_tag', 'access_tag');

-- All queries automatically filtered by tenant
MATCH (n)
RETURN n;  -- Only returns nodes for current tenant

Real-Time Updates

Combine CDC with webhooks for real-time processing:

# cdc-config.yaml
cdc:
  enabled: true
  webhooks:
    - url: "https://analytics.example.com/process"
      events: ["node.created", "edge.created"]

Use embeddings for similarity-based queries:

-- Find similar items
MATCH (item:Product)
WHERE distance(item.embedding, $query_embedding, 'cosine') < 0.5
RETURN item.name, distance(item.embedding, $query_embedding, 'cosine') AS distance
ORDER BY distance ASC  -- cosine distance: lower = more similar
LIMIT 10;

Next Steps