Retail ETL Warehouse
A production-style ETL pipeline: raw retail exports flow through a validated data-quality gate into a star-schema warehouse, SQL marts and a read-only analytics API.
Problem
Retail exports often contain dirty or inconsistent records that silently corrupt downstream reporting if they aren't caught.
Solution
An ETL pipeline where raw exports pass through a validated data-quality gate before loading into a star-schema warehouse, with a read-only analytics API on top.
Overview
A single malformed row in a raw retail export can silently corrupt every downstream report built on top of it — and by the time anyone notices, the bad numbers may already have driven a decision.
How it works
Raw retail exports pass through a validated data-quality gate — schema checks, null checks, business-rule checks — before they’re ever allowed into the warehouse. Only validated data reaches the star-schema warehouse and SQL marts, which are exposed through a read-only analytics API for downstream consumption.
Result
A pipeline where bad data gets caught at the gate, not discovered three reports later.
Further reading
See Why Most Data Portfolios Get Rejected for more on why untested pipelines are a common portfolio failure, and dbt + DuckDB vs a Cloud Warehouse for the reasoning behind this stack’s warehouse choice.
More projects
Fraud Detection System
A classical machine learning system for detecting fraudulent transactions with a focus on precision/recall trade-offs.
Bike-Share Demand Forecasting
Hourly bike-share rental demand forecasting with honest time-series validation, SHAP explainability, and a containerized FastAPI serving layer.
Organizational Network Analysis
Graph-based analysis of organizational networks to surface influence, collaboration and communication patterns.