NeuralBlitz What I do
- Query graph databases efficiently
- Design graph schemas and models
- Traverse complex relationships
- Optimize graph queries
- Build recommendation systems
When to use me
When working with highly connected data, social networks, fraud detection, or recommendation engines.
Graph Concepts
Nodes and Relationships
- Nodes: Entities (people, products, places)
- Relationships: Connections between nodes
- Properties: Key-value pairs on both
- Labels: Node types
- Types: Relationship types
Property Graph Model
CREATE (alice:Person {name: 'Alice'})-[:FRIEND {since: 2020}]->(bob:Person {name: 'Bob'})
Cypher Query Language
Basic Queries
// Match nodes
MATCH (p:Person) WHERE p.name = 'Alice' RETURN p
// Relationships
MATCH (p1:Person)-[:FRIEND]->(p2:Person) RETURN p1, p2
// Pattern matching
MATCH (p:Person)-[:FRIEND]->(friend)-[:FRIEND]->(friendOfFriend)
WHERE p.name = 'Alice'
RETURN friendOfFriend.name
Filtering
MATCH (p:Person)
WHERE p.age > 25 AND p.name STARTS WITH 'A'
RETURN p
Aggregation
MATCH (p:Person)-[:FRIEND]->(friend)
WITH p, count(friend) AS friendCount
WHERE friendCount > 10
RETURN p
Path Finding
// Shortest path
MATCH path = shortestPath((a:Person)-[*]-(b:Person))
WHERE a.name = 'Alice' AND b.name = 'Charlie'
RETURN path
Graph Modeling
Design Principles
- Start with questions, not entities
- Use meaningful relationship types
- Model for queries
- Consider traversal depth
- Denormalize appropriately
Common Patterns
- Friend of Friend: Social connections
- Hierarchies: Org charts, categories
- Sequences: User journeys, events
- Multiple hops: N-degree connections
Use Cases
Social Networks
- Friend recommendations
- Influence analysis
- Community detection
Fraud Detection
- Unusual patterns
- Ring detection
- Connection analysis
Recommendation Engines
- "Users who bought this also bought"
- Skill matching
- Content recommendations
Network Analysis
- IT infrastructure
- Supply chain
- Disease spread
Knowledge Graphs
- Semantic search
- Entity resolution
- Taxonomy
Database Systems
- Neo4j: Most popular, Cypher
- Amazon Neptune: Multi-model (graph + RDF)
- ArangoDB: Multi-model (graph + document)
- Apache Jena: RDF triple store
- TigerGraph: High performance
Performance Optimization
- Indexes on properties
- Relationship density consideration
- Avoid excessive traversal
- Use projections
- Partition large graphs