All the work, with the source of each number.
Each row shows a label. Real is from a live system. Sample is rebuilt with false data. Published is coursework output that you can run again. Modelled is a projection, not a result. Prototype is a working demo with sample data.
- Stay for the exhale
I turned a yoga teacher's classes into a membership product. I did the strategy, the design, the build and the operations. Result: 54 written plans.
- A bank reconciliation that explains each difference
A working sample of the reconciliation method from my years as Operations & Finance Lead. Excel formulas match 119 bank lines to 121 ledger entries, explain each difference and reconcile to 0.00. Result: 15 of 15 planted differences caught, 0.00 difference.
- SQL case studies: revenue share, donor segments and a 3NF schema
Three SQL studies on small course datasets. Each one starts with a business question. Every figure is query output that you can run again. Result: 3 of 14 donors give 66% of gift value.
- Monthly owner report for a three-property portfolio
A rebuilt sample of the monthly report that each asset owner received: occupancy, rent roll, NOI against budget, capital projects and the NOI share of each owner. Result: 14 of 15 units occupied, NOI 7,925 below budget.
- Restaurant margin and guest retention dashboards
Two Tableau dashboards that map each view to a margin, labour or retention decision. The model gives a projected 2% net-margin lift. Projection: 2% net-margin lift from three changes.
- Stella: a guest-inquiry assistant that routes by value
I designed and prototyped an assistant that answers routine questions, books small tables and sends high-value requests to a manager. The demo shows each routing decision. Result: Parties of 6 or more go to a manager.
- What 218,812 Airbnb reviews talk about
A Python check and keyword scan of a large review dataset. It turns free text into a service-quality snapshot, with aggregate counts only. Result: 48.8% of reviews mention location.
- Cinema screen plan with a Monte Carlo simulation
A group assignment. We simulated 100,000 nights of demand for two screen plans. The plan that we proposed gives more expected revenue from the same five screens. Projection: Expected nightly revenue $5,756 → $5,975 (modelled).
- Ticket resale alerts with a classification tree
A group assignment. We trained a classification tree that gives each baseball game a resale alert level: low, medium or high. Result: 82.1% accuracy on the test season, against a 58.9% baseline.
- Baseball attendance model, and the input that made it look too good
A group assignment. We built a regression for attendance at Atlanta Braves home games. One input contained the attendance value, and the fit is lower without it. Result: Fit (R²) 0.991 with the leaked input, about 0.95 without it.
- Win prediction: five points better than a simple guess
A group assignment. We built a logistic regression that predicts a win or a loss for each home game, and compared it with a simple baseline. Result: 56.1% accuracy, against a 51.1% baseline.
- BigQuery SQL on Yelp business and review tables
A group assignment. We answered seven business questions with BigQuery SQL on Yelp business, user, review and tip tables.
- Client intake system: from process map to data design
A systems-analysis design for service intake: capture, validate, classify, route, reply, log and report. It applies course methods to the intake problems that I saw in consulting.
- Player activity data: an individual planning report
My individual plan for a group project on player data from a research game server. I joined two tables, explored the data in three charts and proposed a model. Result: 196 players and 1,535 sessions joined into one table.