DS 4320 Project 1: Building a Relational Movie Recommendation Dataset from MovieLens

This project builds a relational secondary dataset from the MovieLens “latest small” dataset to support personalized movie recommendations. The raw data is transformed into a normalized relational structure with linked tables for users, movies, ratings, tags, links, genres, and movie_genres. Using Python, SQL, DuckDB, and collaborative filtering, the project demonstrates how relational data design and machine learning can be used together to recommend movies a user is likely to enjoy.本项目以MovieLens「最新小型」数据集为数据源，构建了一套关系型次级数据集，用于支撑个性化电影推荐任务。原始数据经转换后形成规范化关系型结构，包含用户、电影、评分、标签、链接、影片类型（genres）以及电影-类型关联表（movie_genres）等关联表。本项目借助Python、SQL、DuckDB与协同过滤（collaborative filtering）技术，演示了如何将关系型数据设计与机器学习相结合，为用户推荐其可能喜爱的电影。