thanhtai435/agentic-bi-ecommerce

该数据集是巴西电子商务平台Olist的数据样本，包含500个订单的Parquet格式数据，完整数据集包含约10万个订单。数据集采用星型模式设计，包含订单、订单项、客户、产品、卖家、支付和评价等表格。该项目是一个端到端的数据工程系统，集成了流式管道、湖仓架构（Medallion）和智能BI，涵盖数据摄取、转换（Bronze → Silver → Gold层）、机器学习（关联规则、聚类、分类）和可视化（Streamlit仪表盘 + 智能BI）等多个阶段。

This dataset is a sample of Brazilian e-commerce platform Olists data, containing 500 orders in Parquet format, with the full dataset comprising approximately 100K orders. The dataset is designed in a star schema, including tables for orders, order items, customers, products, sellers, payments, and reviews. The project is an end-to-end data engineering system integrating streaming pipelines, lakehouse architecture (Medallion), and Agentic BI, covering multiple stages such as data ingestion, transformation (Bronze → Silver → Gold layers), machine learning (association rules, clustering, classification), and visualization (Streamlit dashboard + Agentic BI).