Forecasting with Linear Regression Models

Workshop Overview

Welcome to the Forecasting with Linear Regression Models workshop materials!

This workshop focuses on the foundation of feature-based forecasting with linear regression using R. The workshop is beginner-friendly and covers the following topics:

• How to frame a time series as a regression problem
• Feature engineering techniques for time-aware data
• Modeling trend, seasonality, outliers, and breaks
• Practical tips for evaluation and validation

Latin America 3D map (created with Midjourney)

Workshop Materials

The workshop is organized into 7 notebooks:

📊 Part 1: Data Exploration and Analysis

1. Data - Introduction to the datasets and time series objects
2. Time Series Decomposition - Understanding trend, seasonal, and irregular components
3. Correlation Analysis - Analyzing autocorrelation and partial autocorrelation
4. Seasonal Analysis - Deep dive into seasonal patterns

🎯 Part 2: Forecasting Models

5. Modeling Trend - Building linear and piecewise trend models
6. Modeling Seasonality - Incorporating seasonal patterns and structural breaks
7. Probabilistic Forecast - Creating simulation-based forecasts with uncertainty quantification

Environment Settings

To ensure a smooth workshop experience, you can set up your R environment using one of the following options:

Option 1: Dev Container (Recommended)

If you’re using VSCode, you can use the Dev Container extension to run the workshop in a pre-configured containerized environment:

1. Install Docker Desktop
2. Install the Dev Containers extension in VSCode
3. Open this repository in VSCode
4. When prompted, click “Reopen in Container” or run the command “Dev Containers: Reopen in Container”

The container includes all required R packages and dependencies pre-installed.

Option 2: Using renv

If you prefer to work in your local R environment, you can use renv to install the exact package versions used in the workshop:

1. Ensure you have R (version 4.0 or higher) installed

2. Open R in your preferred IDE (Positron/RStudio/VScode) in the project directory

3. Install renv if not already installed:

   install.packages("renv")

4. Restore the environment using the provided renv.lock file:

   renv::restore()

This will automatically install all required packages with the exact versions specified in the renv.lock file.

About the Workshop

When: December 1st, 2025 from 9am to 11am PST

Where: Online

Registration: Eventbrite

This workshop is for practitioners (data analysts/scientists) who wish to learn how to forecast with regression models. It assumes no background in time series analysis and forecasting, but assumes basic knowledge of probability, linear regression, and R programming.