Model continuous variables to support decisions: developing linear models using Excel

Participants will learn how to examine relationships between variables and build simple predictive models that estimate how a numerical outcome changes in response to one or more influencing factors. The course combines principles with practical applications: deciding when a model is appropriate, how reliable it is, and how its results should be interpreted. The course uses Excel as a familiar environment to implement the concepts.

This course is designed for professionals who want advance beyond data exploration and begin modelling numerical outcomes using Excel.  Through hands‑on exercises, participants will:

• propose models based on observed relationships in data
• estimate and interpret model outputs
• assess whether a model is useful for inference and prediction
• communicate results clearly, including uncertainty and limitations.

Day 1:

Part I

Understanding and modelling relationships

• The line as a model for numerical variables, selecting predictors and understanding how one predictor variable influences the dependent variable
• Exploring relationships between variables: scatter plots,  covariance, and correlation.

Part II

Building a First Predictive Model: creating a model to predict a numerical outcome from one influencing factor

• What tools do we have? An overview of Excel capabilities
• The key assumptions behind developing a linear model and how to verify they are met
• Evaluating how well the model represents the data: goodness of fit.
• How useful is the model for predictions? The concept of statistical inference and significance

Day 2

Part I

Improving and Adapting Models.

• Applying transformations to better represent observed patterns
• Interpreting results after transformation

Part II

Modelling with multiple influencing variables

• Identifying potential issues when predictors are related to each other
• Interpreting the contribution of each variable to the outcome

Comparing between different models to support selection

• Examine relationships between variables to determine whether a linear model is suitable and selecting variables to use in the model
• Evaluate model performance, assess suitability for explanation and prediction
• Explain and verify the key assumptions of a linear model using visual and numerical diagnostics
• Communicate model results, assumptions, and limitations clearly

Classroom delivery coupled with hands-on applications using example data files and purpose-built excel templates.