Building a GLM from Start to Finish

The PCPA project requires candidates to work through a full modeling workflow. This includes exploring data, cleaning and transforming it, building and improving a GLM, interpreting diagnostics, and communicating results in a concise technical summary.

This course was designed based on feedback from PCPA Project beta testers, who recommended a detailed walk-through of building a GLM. They felt a step-by-step tutorial could be especially helpful for candidates pursuing the PCPA, and for those who have never built a model from start to finish.

Follow a guided case study to prepare real insurance data, build a GLM, evaluate it, and communicate results using R, Python, or SAS.

For PCPA Project Candidates

Once you have passed your PCPA exam, this optional course is designed to support your work on the PCPA project. It helps you build a GLM from start to finish, practice the core skills the project evaluates, and reduce uncertainty about what steps to take and in what order.

For Professional Learning GLMs

For professionals seeking practical GLM skills, this course offers a structured, hands-on approach to building a GLM from start to finish in a property and casualty insurance context.

The Course

Module 1: Prepare insurance data for modeling

A guided case study focused on converting datasets into clean, usable modeling data.

~100 minutes of video
Expected learner time: ~3–4 hours including follow-along steps (pause/resume anytime)

The first module of this course will focus on how to convert insurance data sets to usable data for use in GLMs. This is a directed case study module that utilizes data and is designed for the learner to follow along with the instructions and guidance of the videos and texts.

The objective of this course is that you will learn and apply the knowledge and skills to analyze data, convert data to a usable form, and be able to apply these skills to future work.

Module 2: Develop a GLM from start to finish

A second guided case study that walks through GLM construction, evaluation, and communication.

Short lessons designed to support stop/restart based on time availability
Reflection points and (where applicable) knowledge checks

This module is a directed case study designed for active participation by the learner.

The objective of this course is that you will learn and apply the knowledge and skills to develop a GLM and apply the skills to future work.

A note regarding lesson length: This course is designed around the steps needed to build a predictive analytics generalized linear model. While it seems that there are many lessons to complete, most of them are short and have been designed to allow for the course to be stopped and restarted based on time availability and needs of the learner.

Outcomes

Data Preparation

• Convert raw insurance datasets into an analysis-ready form
• Handle common issues: missingness, outliers, transformations, imperfect data

GLM Development

• Select an appropriate GLM type for the outcome
• Build a GLM, evaluate fit and diagnostics, and iterate to improve performance

Interpretation & Communication

• Translate model output into clear technical explanations
• Summarize results for non-technical stakeholders with business implications

Build the skill the PCPA project actually demands: end-to-end modeling

This course is designed as an instruction manual that walks you through the workflow step by step. It helps you avoid stitching together multiple references or guessing the correct order of operations.

Choose your software version

R Version

R + RStudio workflow for data prep and development

Register for the R Version

Python Version

Python tooling and patterns

Register for the Python Version

SAS GLM Course

SAS-specific implentation

Coming Soon

The CAS notes the project code can be submitted in R, Python, or SAS, and references are not intended to signal a preferred language.

Pricing

Pricing that fits your situation

PCPA Candidates (special price)

$50 for active PCPA candidates (Discount code included in your exam pass confirmation)

All other learners

$300 members / $350 non-members

Frequently Asked Question

Is the GLM course a required component of the PCPA project?

No. The GLM course is not a required component of the PCPA project. It is an optional, guided walk-through for candidates who want a structured, start-to-finish build experience to support the GLM portion of the project.

Which language should I pick?

Pick the language you plan to use for your project submission (R, Python, or SAS). The CAS notes the project code can be submitted in R, Python, or SAS, and references are not intended to signal a preferred language.

How long do I have to complete it?

It’s self-paced. Most learners complete the instruction in ~8 hours, plus time to follow along with the case study steps.

Does this replace the PCPA exam readings?

No, this is an application-focused build. It’s meant to complement exam prep and reduce friction when moving from reading to doing.

How do I get the candidate discount?

Upon passing your PCPA Exam, you will receive a confirmation email detailing the next steps for your PCPA project. Your discount coupon code will be included directly within that email. Please ensure you check your inbox (and spam folder) for this notification. If you need further support, contact ACS at acs@casact.org.