Portfolio

Projects

Each one starts from a real question and real data. The self-directed builds are larger and engineered end to end; the case studies take a dataset or a brief from my studies and carry it further.

Self-directed

Built end to end

DatabricksdbtDuckDBPythonSQLStreamlitGitHub Actions

AZIMUTH · Launch Market Intelligence Platform

A BI platform that models the global orbital launch market and compares any rocket against the field. The pipeline refreshes, tests and deploys itself every day.

7,659orbital launches, 1957 to today
2,352historical flights in the simulator backtest
0 to 7phases, each with a build-log post
Project page and build log
Case studies

Analysis on real data

TableauLOD expressionsData cleaning

Four Questions from the Celtics Front Office, Answered in Tableau

Shooting, playmaking, roster age and winning without Tatum, answered on 2025-26 data from four related sources.

Read the case study
Rt-testsRegression

The Column You Test Decides the Answer

Employee attrition in R: three noise columns caught early, a statistically valid wrong answer from testing an ID number, and pay explained to 90.5% by job level.

Read the case study
MySQLCTEsWindow functions

One Hospital Resource Limits Everything: Beds. A SQL Deep Dive

101,766 hospital encounters: the 21% of stays lasting 7+ days consumed 43.5% of all bed-days.

Read the case study
PythonpandasLag analysis

The Best-Controlled Number Tells You the Least

Six months of iron ore flotation data in pandas: a hidden 13-day recording gap, two measurement rates in one table, and a reagent that appears to chase impurity instead of preventing it.

Read the case study
MySQLAggregationsPeriod analysis

What Happened to $1.76 Billion in World Bank Funding? A SQL Analysis of Myanmar

World Bank IDA lending to Myanmar across eras, sectors and individual projects, in MySQL: 11,171 records behind a $1.76 billion story.

Read the case study
TableauKPIs

Massachusetts Education Dashboard

Dropout patterns across 1,861 Massachusetts schools: 37 critical schools and a strong relationship with economic disadvantage.

Read the case study
ExcelCorrelation

What 2,000 Food Delivery Customers Reveal about Spending and Marketing

Income explains 67% of spending variation, and discount-heavy campaigns underperform for high-value segments.

Read the case study