9 Modeling and Model Building

Models are how BI moves from description to prediction and prescription. The previous chapters focused on preparing, exploring, and visualizing data — essential steps, but ultimately preliminary. The value of BI is realized when organizations use data to make better decisions, and models are the tools that make this possible. A model takes the patterns discovered in data and formalizes them into a structure that can explain, predict, or optimize business outcomes.

This chapter covers the principles of model building: what types of models exist, how to select and train them, how to evaluate their performance, and how to deploy them responsibly. We introduce the modeling workflow and demonstrate key concepts with R code, preparing you for the hands-on regression modeling in Chapter 10.

Chapter Goals

Upon concluding this chapter, readers will be able to:

1. Distinguish among descriptive, predictive, and prescriptive models and identify when each is appropriate.
2. Describe the modeling workflow from problem definition through deployment and monitoring.
3. Apply the principles of parsimony and generalizability when selecting a model, and explain why overfitting is a risk.
4. Evaluate model performance using in-sample and out-of-sample metrics, including train/test splits and cross-validation.
5. Explain how AI and AutoML tools are changing the modeling process, and why human judgment remains essential.