Projects
Every card is labeled by status: Complete, In Development, Architecture Complete, or Planned Prototype. A project can appear under more than one category when it genuinely spans both.
PostgreSQL analysis for business performance and decision support — data validation, customer segmentation, trend queries, and ranking logic using CTEs and window functions.
Business Challenge
Enterprise sales & operations reporting is only as trustworthy as the data model underneath it — most reporting projects jump straight to dashboards without validating that foundation first.
Solution
A configuration-driven, validated data platform (Kimball-modeled) built to be trustworthy before any dashboard sits on top of it — foundation-first, not dashboard-first.
Business Outcome
In Development — the validated data foundation is built; the goal is a platform other reports (like the Executive Sales Performance Dashboard below) can build on without re-validating the data every time.
A configuration-driven enterprise sales analytics platform built on a large-scale synthetic B2B manufacturing dataset, using Kimball dimensional modeling. Completed: the dataset architecture, dimensional model, validation framework, documentation, and supporting BI reporting work. The broader platform — the full semantic and reporting layer across all of it — remains in development.
Technologies Used
Business Challenge
Leadership needed a trustworthy, self-service view of sales and profitability performance without waiting on a manual report cycle.
Solution
An interactive Power BI report with a reusable DAX measure library so every page uses the same KPI definitions, plus drill-through from summary KPIs to transaction-level detail.
Business Outcome
A single source of truth for sales performance — no reconciling numbers that differ depending on which report or analyst produced them.
A finished, standalone Power BI deliverable built from the ESIP dataset — sales, profitability, product, and regional performance with a reusable DAX measure library, executive KPI cards, and drill-through analysis. Distinct from the broader ESIP platform above, which is still in development as a whole.
Technologies Used
Originally developed as the Grace & Stella Commercial Intelligence Platform.
Business Challenge
Commercial and leadership teams needed to explore revenue, profitability, and promotional performance themselves, not wait on a static export.
Solution
A live, interactive Streamlit application (Python/Pandas/Plotly) covering revenue, margin, promotions, channels, products, and country performance.
Business Outcome
A deployed application stakeholders can open and explore directly — real evidence of shipped, usable output, not a screenshot.
An executive dashboard analyzing revenue, gross profit, promotions, channels, products, and country performance — deployed as a live, interactive Streamlit application using portfolio-safe synthetic data. No private client, customer, or employer information is displayed.
Technologies Used
Business Challenge
Enterprise sales & operations reporting is only as trustworthy as the data model underneath it — most reporting projects jump straight to dashboards without validating that foundation first.
Solution
A configuration-driven, validated data platform (Kimball-modeled) built to be trustworthy before any dashboard sits on top of it — foundation-first, not dashboard-first.
Business Outcome
In Development — the validated data foundation is built; the goal is a platform other reports (like the Executive Sales Performance Dashboard below) can build on without re-validating the data every time.
A configuration-driven enterprise sales analytics platform built on a large-scale synthetic B2B manufacturing dataset, using Kimball dimensional modeling. Completed: the dataset architecture, dimensional model, validation framework, documentation, and supporting BI reporting work. The broader platform — the full semantic and reporting layer across all of it — remains in development.
Technologies Used
A conceptual architecture — not a deployed or production solution — for an enterprise data-quality review platform built entirely on the Microsoft Power Platform, connected to Python validation services already proven in my other projects.
Prototype planned next.
Business Challenge
Manually identifying supplier products from photos and drafting listings is slow, inconsistent, and doesn't scale.
Solution
A human-in-the-loop AI-assisted design — AI accelerates identification and drafting, a person makes every publish decision — with full audit history built into the architecture from the start.
Business Outcome
Architecture Complete — the design de-risks the build phase before a line of code is written; no application exists yet, so there's no production outcome to report.
A designed (not yet built) system for AI-assisted product enrichment from supplier photos, with a human-in-the-loop review and approval workflow, product governance, and full audit history. Solution architecture, data model, and integration design are complete; a mocked-data prototype is the next planned phase — no application code exists yet.
Technologies Used
A conceptual architecture — not a deployed or production solution — for an enterprise data-quality review platform built entirely on the Microsoft Power Platform, connected to Python validation services already proven in my other projects.
Business Challenge
Leadership needed a trustworthy, self-service view of sales and profitability performance without waiting on a manual report cycle.
Solution
An interactive Power BI report with a reusable DAX measure library so every page uses the same KPI definitions, plus drill-through from summary KPIs to transaction-level detail.
Business Outcome
A single source of truth for sales performance — no reconciling numbers that differ depending on which report or analyst produced them.
A finished, standalone Power BI deliverable built from the ESIP dataset — sales, profitability, product, and regional performance with a reusable DAX measure library, executive KPI cards, and drill-through analysis. Distinct from the broader ESIP platform above, which is still in development as a whole.
Technologies Used
Originally developed as the Grace & Stella Commercial Intelligence Platform.
Business Challenge
Commercial and leadership teams needed to explore revenue, profitability, and promotional performance themselves, not wait on a static export.
Solution
A live, interactive Streamlit application (Python/Pandas/Plotly) covering revenue, margin, promotions, channels, products, and country performance.
Business Outcome
A deployed application stakeholders can open and explore directly — real evidence of shipped, usable output, not a screenshot.
An executive dashboard analyzing revenue, gross profit, promotions, channels, products, and country performance — deployed as a live, interactive Streamlit application using portfolio-safe synthetic data. No private client, customer, or employer information is displayed.
Technologies Used
Automated financial reconciliation and balance-monitoring system for patient-ledger monitoring, revenue reconciliation, adjustments, and exception review.
A governed ETL pipeline (Google Sheets + Apps Script) moving synthetic commerce data through validation into clean analytical tables, deterministic KPI aggregation, and a 5-page interactive dashboard. 43 of 43 automated tests passing.
PostgreSQL analysis for business performance and decision support — data validation, customer segmentation, trend queries, and ranking logic using CTEs and window functions.
Business Challenge
Enterprise sales & operations reporting is only as trustworthy as the data model underneath it — most reporting projects jump straight to dashboards without validating that foundation first.
Solution
A configuration-driven, validated data platform (Kimball-modeled) built to be trustworthy before any dashboard sits on top of it — foundation-first, not dashboard-first.
Business Outcome
In Development — the validated data foundation is built; the goal is a platform other reports (like the Executive Sales Performance Dashboard below) can build on without re-validating the data every time.
A configuration-driven enterprise sales analytics platform built on a large-scale synthetic B2B manufacturing dataset, using Kimball dimensional modeling. Completed: the dataset architecture, dimensional model, validation framework, documentation, and supporting BI reporting work. The broader platform — the full semantic and reporting layer across all of it — remains in development.
Technologies Used
Prototype planned next.
Business Challenge
Manually identifying supplier products from photos and drafting listings is slow, inconsistent, and doesn't scale.
Solution
A human-in-the-loop AI-assisted design — AI accelerates identification and drafting, a person makes every publish decision — with full audit history built into the architecture from the start.
Business Outcome
Architecture Complete — the design de-risks the build phase before a line of code is written; no application exists yet, so there's no production outcome to report.
A designed (not yet built) system for AI-assisted product enrichment from supplier photos, with a human-in-the-loop review and approval workflow, product governance, and full audit history. Solution architecture, data model, and integration design are complete; a mocked-data prototype is the next planned phase — no application code exists yet.
Technologies Used
A conceptual architecture — not a deployed or production solution — for an enterprise data-quality review platform built entirely on the Microsoft Power Platform, connected to Python validation services already proven in my other projects.
A governed ETL pipeline (Google Sheets + Apps Script) moving synthetic commerce data through validation into clean analytical tables, deterministic KPI aggregation, and a 5-page interactive dashboard. 43 of 43 automated tests passing.