Building an AI Powered Video Recommender - Knowledge Graphs, NLP, NetworkX Tutorial

This tutorial guides you through building an AI-powered video recommendation system using knowledge graphs, natural language processing (NLP), and the NetworkX library. Learn how to represent educational videos as nodes in a graph with relationships captured by edges. Process video metadata (titles, descriptions, durations) using spaCy for key concept extraction and SentenceTransformer for generating semantic embeddings that calculate similarity scores between videos. Construct a knowledge graph with NetworkX where edges represent both semantic similarity and logical prerequisites, while nodes store attributes like difficulty level, topics covered, and duration. Explore learning paths, topic relationships, and video prerequisites through various analyses and visualizations using NetworkX and PyVis. The tutorial also demonstrates how to serialize the knowledge graph into a SQLite database with separate tables for nodes, edges, and embeddings, allowing you to learn data persistence techniques. By the end, gain hands-on experience with building sophisticated recommendation systems using graph theory and NLP, with practical knowledge of storing, querying, and manipulating graph data. All code is available on GitHub, along with the input CSV file needed for implementation.