This script shows how to classify fraudulent orders by preprocessing (checking for correlation and zero/near-zero variance), selecting features (logistic regression), and using R's CARET package to train several machine learning models (recursive partitioning, logistic regression, random forest, neural network, and support vector machine).

This script shows how I segment my company's customer base by clustering average order values.

This script shows how I loop through three types of forecasting models in order to predict unit sales for over 850 products. The three models are random walk, the entire family of exponential smoothing models, and multiple ARIMA models.