# Graph Reasoning and Analytics

# Graph Reasoning and Analytics

Graph reasoning combines the structural power of the knowledge graph with advanced algorithms and LLMs to derive deep insights from your data.

# Graph Algorithms

Curiosity Workspace provides built-in support for common graph algorithms to analyze the structure of your data.

# Centrality

Identify the most important or influential nodes in your graph.

• PageRank: Measures the relative importance of nodes based on their connections.
• Betweenness Centrality: Finds nodes that act as bridges between different communities.

# Community Detection

Group nodes that are more densely connected to each other than to the rest of the graph.

• Louvain Method: An efficient algorithm for finding communities in large graphs.
• Label Propagation: A fast method for identifying clusters based on node labels.

# Pattern Matching

Use pattern matching to find specific structural arrangements in the graph, such as cycles, stars, or triangles. This is essential for detecting anomalies and recurring domain patterns.

# AI-Driven Graph Reasoning

Combine graph algorithms with LLMs for sophisticated analysis:

1. Contextual Retrieval: Use graph algorithms to identify the most relevant neighborhood for a user's question.
2. Reasoning: Feed the structural metadata and node properties into an LLM.
3. Synthesis: The LLM explains the relationship or identifies the underlying cause of a pattern.

# Analytic Queries

Run aggregations across graph structures to answer complex business questions:

• "What is the average number of connections for nodes of type 'Customer'?"
• "Which 'Project' nodes have the highest growth in related 'Incident' nodes over the last 30 days?"

# Next Steps

• Learn about LLM Agents
• Explore Custom Queries