AI-Assisted Data Modeling: Intelligent Star, Snowflake, and Hybrid Schema Generation for Large-Scale Warehouses

Abstract

Designing efficient data warehouse schemas—such as star, snowflake, and hybrid models—has traditionally been a manual, expertise-driven process requiring deep knowledge of business processes, data dependencies, and query performance optimization. With the exponential growth of data sources and the increasing shift toward real-time analytics, traditional modeling methodologies face limitations in scalability, accuracy, and development speed. This research introduces an AI-Assisted Data Modeling Framework that automates the design of warehouse schemas using clustering algorithms, NLP-assisted semantic understanding, machine learning–based pattern identification, and rule-based optimizations. The proposed system intelligently extracts metadata, identifies facts and dimensions, detects hierarchies, and selects optimal schema types based on analytical workloads, data cardinality, and normalization requirements. A detailed case study on a retail enterprise demonstrates improvements in schema-design time, structural accuracy, dimensional hierarchy detection, and performance predictions. This work follows a structured academic format inspired by the style and clarity of your sample research paper on blockchain-enabled AI systems.