Inventory and operations analytics — stock, pricing, demand, and reporting. Retail, supply chain, wholesale distribution, and manufacturing — built on BigQuery, SQL, Power BI, and DAX.

• BI & Analytics: Power BI, Excel, Power Query (M)
• Languages & Databases: SQL, DAX, R, Google BigQuery, MATLAB
• Tools & Version Control: Git, GitHub, Notion

Tech Stack: Google BigQuery, SQL, Power BI, DAX Focus: Exit timing, size-curve scarcity, demand risk scoring

99,956 real StockX transactions (2017–2019) across two brands running opposite playbooks — scarcity vs volume — modeled through a BigQuery staging pipeline and a Power BI star schema, with every headline number reconciled across three independent paths (CSV profile, SQL, report).

• Diagnosed a uniform whitespace defect silently breaking 72% of brand joins, and validated the pipeline end-to-end with five checks plus multi-path reconciliation of all published figures.
• Showed the market's apparent premium collapse after year one is a composition artifact: split by brand, aged Off-Whites appreciate to a 578% median premium while Yeezys decay to 31% — making exit timing a brand-specific decision.
• Scored all 50 models on price volatility (CV) vs premium to map hero SKUs and demand risk, and quantified size-curve scarcity: tail sizes carry 3–4x core-size premiums, with size 15 running 100% Off-White.

👉 View the Full Project

Tech Stack: Power BI, DAX, Power Query (M), Excel Focus: SKU profitability, cash velocity, capital payback

A 3-year SKU-level profitability model across 74 assets and 12 selling seasons — star schema, class-aware DAX, and a disconnected 12-period timeline replacing standard Time Intelligence in a non-calendar seasonal model.

• Modeled daily cash velocity across the full catalog, isolating a ~95x spread between the fastest and slowest cash-positive SKUs ($17.11/day vs $0.18/day).
• Engineered class-aware ROI and break-even logic across 4 asset classes, surfacing a 1,048% maximum 3-year ROI and exactly one structurally unprofitable SKU flagged for delisting.
• Mapped capital payback on a 12-period cumulative curve: tree-class assets run Year-1 deficits, break even in Fall of Year 2, and lead all classes by Year 3 — full-catalog 3-year net position of $57,564.

👉 View the Full Project

• College Cost vs. Earnings — R linear regression on U.S. Dept. of Education data; every $1 in tuition correlated with $4.59 in 10-year earnings (R² = 0.344)
• Used Car Price Valuation — R regression comparison; improved R² from 0.33 to 0.94 by adding categorical features
• Wheatfield Harvest Simulation — MATLAB stochastic simulation, 3,000 runs, modeling yield variability under random weather

• LinkedIn