Data Analyst and Data Engineer with 9 years of domain expertise in international trade, logistics, and supply chain operations. I build end-to-end analytical systems — from raw operational data to executive dashboards — using SQL Server, Python, and Power BI.

My work lives at the intersection of data engineering and logistics analytics: medallion-architecture warehouses, predictive delay models, and KPI frameworks that turn fleet, freight, and customs data into operational decisions.

Category	Technologies
Databases	SQL Server T-SQL
Languages	Python TypeScript HTML CSS
Data & ML	Pandas NumPy scikit-learn XGBoost Matplotlib
BI & Viz	Jupyter
Architecture	Medallion (Bronze → Silver → Gold) Star Schema ETL/ELT Data Quality
Tools	Git Docker Playwright Next.js

Production-grade data warehouse for a 120-truck, 85,000-load/year logistics operation.

549,706 rows → Bronze → Silver → Gold → 11 business views → $298.6M revenue explained

Complete medallion pipeline on SQL Server 2025: 14 operational tables transformed into a conformed star schema with surrogate keys, enforced referential integrity, and a management-queryable semantic layer. Every KPI cross-validated against source data — 0 rows dropped, 0 orphaned references.

→ Repository | → Documentation

Full medallion-architecture warehouse integrating CRM and ERP source systems.

Bronze → Silver → Gold pipeline built entirely in T-SQL with stored procedures, surrogate keys via ROW_NUMBER(), MDM business rules with CRM/ERP fallback logic, and comprehensive data quality tests. Draw.io architecture diagrams document every layer.

→ Repository

Maritime AIS data analysis — 10.6 million vessel tracks across six US regions in 2025.

GeoPackage spatial database profiled with SQL and Python: fleet composition (45% recreational, 26% commercial), 2.6× seasonal amplitude (August peak vs December trough), regional traffic distribution, and vessel-type behavior patterns. Full case study with SQL queries, EDA charts, and maritime planning recommendations.

→ Repository | → Case Study

Predictive modeling of supply chain delays with 84.2% accuracy.

Polyglot analysis across Excel, SQL Server, and Python/Jupyter. Feature-engineered 15 operational variables, trained XGBoost/Random Forest/Gradient Boosting classifiers, identified traffic and waiting time as top delay predictors. Packaged with an interactive Excel dashboard and a standalone SQL analysis script.

→ Repository

Medallion-architecture analysis of GHG emissions across 1,016 industries.

End-to-end pipeline: raw CSV → Bronze → Silver → Gold views on SQL Server → Python EDA with PyODBC → Power BI dashboard. Right-skewed distribution analysis, outlier detection, industry ranking with actionable sustainability recommendations.

→ Repository

Realistic ocean freight data generator for data engineering workloads.

862-line Python generator producing relational shipping datasets — ports, vessels, shipments, containers, bills of lading, tracking events — with 95% clean / 5% anomalous records, weighted trade lanes, and peak-season simulation. Outputs Parquet or CSV at scale (tested to 100M+ shipments).

→ Repository

Playwright-powered web scraping toolkit for customs tariff and legislation data.

Modular scraper architecture extracting HS code classifications from public Egyptian Customs pages with legal disclaimers, sample output, and test suite.

→ Repository

• Data Warehousing — designing medallion-architecture warehouses on SQL Server with enforced data quality and audit trails
• Logistics Analytics — building KPI frameworks, delay prediction models, and operational dashboards for fleet, freight, and supply chain
• Pipeline Engineering — idempotent ETL/ELT pipelines with TRY_CONVERT-safe casting, orphan handling, and row-level auditability

I spent nearly a decade inside international trade operations — customs clearance, fleet coordination, exception resolution, freight forwarding — before pivoting into data. That domain fluency is my edge: I do not just model logistics data; I understand what a detention hotspot costs, why 44.6% on-time delivery for three straight years signals a structural problem, and which questions management will ask before they ask them.

Every project in this profile combines both halves: the operational reality of moving goods across borders and the engineering discipline of production-grade data systems.