Analytics Portfolio
End-to-end data systems
Each project is a complete data platform: raw data → ingestion → transformation → BI dashboard → AI analyst. Built to showcase real engineering decisions, not tutorials.
Space Financial
2.37M
Rows
31
dbt Models
A regional bank analytics platform with 2.37M rows of synthetic data, 31 dbt models, a live BigQuery warehouse, and an AI analyst built on LangGraph + Claude.
EverMart Retail
Retail analytics for a fictional sportswear company. 50 stores, 1M+ orders, Spark ingestion, dbt marts, AI-driven basket analysis.
In development — check back soon.
How it's built
Data Layer
Python generators create synthetic data. BigQuery hosts the raw, staging, and mart layers. dbt handles all transformations with full tests.
Dashboard Layer
Next.js API routes query BigQuery server-side. Recharts renders the data. Filters hit the API with params — no frontend state except UI.
AI Layer
Claude writes SQL via tool use. A safety layer validates every query before execution. Results stream back to the browser and render as charts inline.