> ## Documentation Index
> Fetch the complete documentation index at: https://fusiondocs.finchai.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Exploring Knowledge Graphs

> Visualize and explore relationships between entities in your research

# Exploring Knowledge Graphs

Knowledge graphs provide a visual way to explore relationships between entities in your documents. See how companies, people, and concepts connect to each other.

## What is a Knowledge Graph?

A knowledge graph is a visual network that shows:

* **Nodes** - Entities (companies, people, topics)
* **Edges** - Relationships between entities
* **Properties** - Additional information about entities

Knowledge graphs help you discover connections that might not be obvious from reading documents linearly.

## Accessing Knowledge Graphs

### From the Sidebar

1. Look for the **Graphs** section in the project sidebar
2. View available graphs for your project
3. Click a graph to open it

### From Graph Mode

1. Switch to **Graph** mode in the chat panel
2. Ask relationship questions
3. View graph visualizations in the response

## Graph Visualization

### Reading the Graph

| Element       | Meaning                                      |
| ------------- | -------------------------------------------- |
| Circles/Nodes | Entities (companies, people, etc.)           |
| Lines/Edges   | Relationships between entities               |
| Node size     | Relative importance or number of connections |
| Node color    | Entity type                                  |
| Edge labels   | Relationship type                            |

### Interacting with the Graph

**Navigation:**

* Click and drag to pan the view
* Scroll to zoom in/out
* Click nodes to select them

**Exploration:**

* Click a node to see its details
* View connections from that node
* Expand to see related entities

## Building Knowledge Graphs

### Automatic Generation

Knowledge graphs are built from:

* Entities recognized in your documents
* Relationships extracted from text
* Connections inferred from context

### Creating a Graph

To generate a knowledge graph for your project:

1. Ensure documents are uploaded and processed
2. Access graph creation from the Graphs section
3. Wait for processing to complete
4. View your new graph

<Info>
  Graph generation requires processed documents with recognized entities. More documents generally produce richer graphs.
</Info>

## Using Graph Mode

Graph mode in the chat panel is optimized for graph exploration.

### Example Queries

**Finding connections:**

* "How is Apple connected to Samsung?"
* "Show me the relationship between these executives"
* "What companies are linked through this technology?"

**Exploring entities:**

* "Show me all companies in this industry"
* "Who are the key people in these documents?"
* "What topics connect these organizations?"

**Path discovery:**

* "What's the connection path between Company A and Company B?"
* "How are these two events related?"

### Interpreting Results

Graph mode returns:

* Visual graph fragments
* Relationship explanations
* Entity details
* Connection paths

## Understanding Relationships

### Relationship Types

Common relationships in knowledge graphs:

| Relationship | Example                           |
| ------------ | --------------------------------- |
| Competition  | Company A competes with Company B |
| Partnership  | Company A partners with Company B |
| Investment   | Investor X invested in Company Y  |
| Employment   | Person X works for Company Y      |
| Acquisition  | Company A acquired Company B      |
| Supply       | Company A supplies to Company B   |

### Relationship Direction

Some relationships are directional:

* A → B (A acquires B)
* A ↔ B (A and B are competitors)

The graph visualization shows direction through arrow orientation.

## Practical Applications

### Competitive Analysis

Use graphs to:

* Map competitive landscapes
* Identify market players
* Understand industry structure

### Due Diligence

Graph mode helps discover:

* Corporate connections
* Key personnel relationships
* Hidden affiliations

### Research Exploration

Find:

* Related companies you weren't aware of
* Connections between topics
* Patterns across your documents

## Graph Features

### Filtering

Narrow your view:

* Filter by entity type
* Show only certain relationship types
* Focus on specific time periods

### Expanding Nodes

Dive deeper:

1. Click a node to select it
2. Use expand options to show connected entities
3. Continue expanding to explore the network

### Layout Options

Adjust the visualization:

* Automatic layout algorithms
* Manual node positioning
* Zoom to fit all nodes

## Saving and Sharing

### Saving Graphs

Your graphs are automatically saved with your project. Access them anytime from the Graphs section.

### Exporting Insights

While the interactive graph stays in the platform, you can:

* Take screenshots for presentations
* Note key findings for reports
* Use insights in report generation

## Combining Graphs with Reports

Use graph insights to enhance your reports:

1. Explore relationships in Graph mode
2. Identify key connections
3. Reference findings in your report generation prompt
4. The AI incorporates relationship insights

<Tip>
  Before generating a report, use Graph mode to understand entity relationships. This context improves report quality.
</Tip>

## Graph Accuracy

### What Graphs Capture

Graphs reflect:

* Entities found in your documents
* Relationships explicitly stated or strongly implied
* Connections across multiple documents

### Limitations

Graphs may not include:

* Relationships not mentioned in documents
* Very subtle or implied connections
* Entities not recognized by the system

### Improving Graph Quality

For better graphs:

* Upload comprehensive source documents
* Include documents that describe relationships
* Use documents with clear entity references

## Troubleshooting

<AccordionGroup>
  <Accordion title="No graphs available for my project">
    Graphs need to be generated from processed documents. Check that you have documents uploaded and processed, then request graph creation.
  </Accordion>

  <Accordion title="Graph is too crowded to read">
    Use filtering to reduce visible entities. Zoom in on specific areas. Focus on a subset of entity types.
  </Accordion>

  <Accordion title="Expected relationship isn't shown">
    The relationship may not be explicitly stated in your documents, or the entities weren't recognized. Check if source documents mention the connection.
  </Accordion>

  <Accordion title="Graph seems incomplete">
    Graphs are built from available documents. Adding more comprehensive source materials will enrich the graph.
  </Accordion>
</AccordionGroup>