About This Architecture

DataSense Advisor ER diagram models a relational schema for an intelligent data preparation platform with User, Dataset, Cleaned Dataset, EDA Module, Data Cleaning Module, Model Recommendation, and Enhancement Module entities. Users own datasets which flow through exploratory data analysis, automated cleaning with configurable strategies (missing value handling, encoding, scaling, outlier detection), and model recommendation based on cleaned data quality. The schema enforces referential integrity via foreign keys and supports JSON-rich attributes for flexible metadata storage, enabling practitioners to track data lineage, cleaning provenance, and ML readiness scores. Fork and customize this diagram on Diagrams.so to adapt the schema for your AWS RDS or DynamoDB backend, or extend it with additional analysis modules. The get_active_df() function bridges modules for real-time dataset access across the pipeline.

People also ask

How should I design a relational database schema for an automated data preparation and ML recommendation platform?

The DataSense Advisor ER diagram provides a production-ready schema with User ownership, Dataset versioning, Cleaned Dataset tracking, and modular analysis entities (EDA, Data Cleaning, Model Recommendation, Enhancement). Foreign key relationships enforce data lineage, while JSON attributes store flexible metadata like cleaning steps, suggested models, and readiness scores.