Airline Graph-RAG Assistant

Airline Graph-RAG Assistant is a Streamlit application that combines airline survey analytics stored in Neo4j with optional semantic retrieval using embeddings.

Airline survey data is easier to explore when structured relationships and relevant semantic context can be retrieved together without allowing generated answers to drift beyond the underlying data.

• Built a Python importer that turns airline survey records into a Neo4j knowledge graph.
• Added graph-backed analytics and optional semantic retrieval with cached embeddings.
• Created a Streamlit interface for asking questions and exploring grounded results.
• Kept external language models optional so the application can run in retrieval-only mode.

1. 01

   Import survey data

   The importer reads airline survey records and creates the corresponding entities and relationships in Neo4j.

2. 02

   Query the graph

   A question is matched to graph-backed analytics over flights, routes, fleets, and survey responses.

3. 03

   Retrieve semantic context

   When enabled, embeddings retrieve additional relevant flight data for the question.

4. 04

   Present grounded results

   The Streamlit interface displays an answer grounded in retrieved graph results, with external language models remaining optional.

• Neo4j knowledge graph
• Airline survey analytics
• Optional semantic retrieval
• Cached embeddings
• Retrieval-only mode
• Streamlit interface