DATA ANALYSIS

Machine Learning

Link

Wine Quality Prediction

Machine learning model to predict the quality of wine samples.

Objective: correctly predict the quality of Portuguese wine based on the physical and chemical characteristics of a sample.

Method: K-Nearest Neighbor classifier model.

Result: 80% accuracy in predicting high quality wines.

Role: School Project

Skills: Python, scikit-learn

Link

Risk Analysis for Mortgage Arrears

Predicting risk for mortgage payment defaults.

Objective: detect tenants at most likely to default on their mortgage payments based on past payments.

Method: K-Nearest Neighbor, Random Forest, and Logistic Regression models.

Result: 77% accuracy for medium-risk defaults, and 58% accuracy for high-risk defaults.

Role: School Project

Skills: Python, scikit-learn

Supply Chain & Logstics

Link

Demand Forecast Planning

Computing demand forecasts for fast-food ingredients.

Objective: anticipate demand for ingredients that will be needed next year.

Method: ARIMA and Holt-Winters models

Result: 75% increase in accuracy with revised seasonality models.

Role: School Project

Skills: Python, scikit-learn, statsmodels

Link

Sourcing & Production Strategy

Supply chain analysis and strategy for winter sport textiles.

Objective: broad goal of creating recommendations based on analysis.

Method: Newsvendor model.

Result: identified new suppliers and production strategy.

Role: School Project

Skills: AMPL

Geopolitical Risk

Link

Political Risk and Microeconomic Development

Measuring the microeconomic impact of political instability on tourism enterprise growth.

Objective: detect tenants most likely to default on their mortgage payments based on past payments.

Method: multivariate regression model.

Result: statistically significant results showing the Sub-Saharan tourism industry faced 2% slower growth than other industries in volatile political climates.

Role: School Paper

Skills: Stata

Google Sites

Report abuse