Publication Date
4-2025
Presentation Length
Poster/Gallery presentation
College
College of Sciences & Mathematics
Department
Math and Computer Science, Department of
Student Level
Undergraduate
SPARK Category
Knowledge
SPARK Session
Independent Presentation
Presentation Type
Poster
Summary
In the era of abundant streaming content, viewers often face choice overload and inefficient recommendation systems limited to isolated platforms. MovieQueue is a graph-based, cross-platform movie recommender system built on a Neo4j knowledge graph and integrated with an interactive Streamlit application. Leveraging both user-specific ratings and extensive movie metadata — including genres, cast, crew, runtime, release year, and audience engagement — the system provides tailored recommendations grounded in content similarity and social context. Users receive recommendations not only based on genre overlap but also through shared actors, directors, and composers with previously highly-rated films, alongside filters for explicit genre selection. A weighted scoring model prioritizes factors such as shared collaborators, movie popularity (vote count), and average rating to deliver more meaningful suggestions. Explanations accompany each recommendation to improve transparency and user trust. The system aims to address the limitations of conventional recommender systems by emphasizing user control, explainability, and rich relational data. MovieQueue demonstrates the potential of graph databases in recommendation tasks and opens new avenues for research into user-centric, explainable AI in media consumption.
Recommended Citation
Juricic, Ryan C., "MovieQueue: Leveraging Graph Databases to Build Recommender Systems" (2025). SPARK Symposium Presentations. 114.
https://repository.belmont.edu/spark_presentations/114
