Realtime Fraud Detection Data Layer
Shaped transaction and feature tables to support high-speed fraud scoring and downstream monitoring.
02 / Skill · SQL
SQL.
I design reliable data models and write production-grade SQL that turns messy operations into trusted business metrics.
About
SQL is the core of my analytics engineering work: Regardless of project there's typically an element of SQL whether it's for use at work to follow the webstore's performance, explore churn or customer segmentation, or on my projects outside of work like using Supabase SQL to set up, monitor, and explore the results from my simulated e-commerce store.
Impact Highlights
- Segmented real guest transactions at ZooTampa with RFM analysis, producing business-ready customer groupings for targeted decision-making.
- Built medallion-style transformation paths for Owl Park — bronze ingestion, silver cleansing, gold dimensional models — that support reusable BI reporting.
- Validate and troubleshoot production point-of-sale data daily, keeping reporting workflows stable for business operations.
- Strengthened KPI consistency by anchoring dashboards to modeled facts and dimensions.
Projects
RFM Customer Segmentation
RFM-based customer segmentation built for the Drop at ZooTampa campaign, identifying high-value and at-risk cohorts from transaction history.
Owl Park Medallion Pipelines
Designed SQL-centric transformation flow from operational Supabase tables into analytics-ready layers powering Power BI.
Dynamic Pricing Simulator Monitoring
Built the aggregate and historical context layer supporting agent pricing and inventory decisions.
Tools
Microsoft Fabric SQL
Databricks SQL
Supabase SQL
MSSQL for Dynamics CRM
DuckDB
Process
- Identify business requirements and define target metrics (OKR, KPIs, supporting metrics).
- Model entities and relationships for stable downstream use.
- Build transformations in audited, testable SQL steps.
- Validate outputs against source totals and edge cases.
- Document assumptions so BI and ML layers stay aligned.