This is a sanitized professional case study. No proprietary datasets, customer information, confidential business rules, internal screenshots, or production code are published here.
The business problem
The business needed a trusted way to explain why packages failed automated processing, determine the underlying cause of USPS assessments, and distinguish valid charges from situations that could be challenged or waived.
My role & analytical ownership
I built the core analytical framework and SQL logic in Databricks, developing causal classifications and business rules that reconcile package activity with assessment outcomes. The work is ongoing, and much of the framework has been in active use since April 2026.
Building trusted logic
- Identify recurring failure patterns and translate them into causal classifications that explain why packages were assessed or failed processing.
- Work with business stakeholders and USPS counterparts to validate the logic against actual processing behavior.
- Refine classifications and business rules based on testing, exceptions, operational feedback, and external validation.
- Use the approved logic to support assessment review and dispute activity where the evidence does not support a charge.
Power BI & future-state architecture
I designed the Power BI reporting approach to provide visibility into assessment trends, root causes, financial exposure, and package-level detail. Today, the logic runs as Databricks SQL and the exported results are connected to Power BI for development and validation. The approved future-state design moves the logic into the enterprise Data Lake, where Power BI will connect directly to a governed materialized view/object.
Approximate reduction in USPS assessment charges since the framework has been in effect.
Why the result matters
The value comes from trust in the analytical logic. USPS has accepted the logic as valid in cases where charges can be waived, giving the business a defensible analytical basis for reducing unsupported assessment charges while also improving visibility into the underlying processing failures.
What this demonstrates
Senior analytical ownership: defining logic, validating it with internal and external stakeholders, designing the BI experience, connecting analytics to financial outcomes, and evolving a working prototype toward a governed enterprise data product.