Actuarial R Tutorials is a worked note on taking a generalized linear model fitted in R and deploying it as an Excel formula. The tutorial, Excelerating the Production Pipeline, writes the coefficients and the R predictions into a workbook, then rebuilds the mean in a spreadsheet cell so the same numbers can be checked without opening R.

The notes are for actuaries, exam candidates, and analysts who estimate in R and review in Excel. The example is deliberately small: petal width on the iris data, with a Gaussian family, a log link, and one log-transformed covariate. The four steps are the same ones you would use for a pricing or reserving GLM.

Open the rendered tutorial below, or rerun Convert R GLM to Excel.Rmd. Load tidyverse, broom, openxlsx, and kableExtra; fit glm(Petal.Width ~ Sepal.Width + log(Petal.Length), family = gaussian(link = "log")); write the coefficient table and predict(..., type = "response") into R GLM in Excel.xlsx; then enter =EXP(coefficients!$B$2+coefficients!$B$3*[@[Sepal.Width]])*[@[Petal.Length]]^coefficients!$B$4 and confirm the new column matches predicted_petal_width.

That formula is the log link solved for the mean. From log(Y) = β0 + β1 X1 + β2 log(X2) the cell is exp(β0 + β1 X1) × X2^β2, and each beta is read from the coefficients sheet. On this fit those values are about −1.835, 0.053, and 1.369. Keeping them in cells means a refit updates the workbook without retyping constants. The same export covers other glm families; a logistic model uses family = binomial(link = "logit").

Sam Castillo wrote the tutorial in 2019 while studying mathematics at UMass Amherst. These pages keep that note — the R Markdown, the notebook render, the sample workbook, and the Excel screenshots — and add a front door on GitHub Pages.

What you can take from it

• A four-step path from an R glm to a checked Excel formula.
• A coefficient sheet so production numbers stay tied to the latest fit.
• Log-link algebra reduced to EXP and a power for log-transformed inputs.
• A published iris example, with the estimates stored beside the R predictions.
• The same export pattern for other GLM families, including logistic regression.
• The prettydoc page, the R notebook, the .Rmd source, and R GLM in Excel.xlsx.

Open the tutorial

Excelerating the Production Pipeline

HTML

Tutorial

Rendered prettydoc of the four steps, the coefficient table, the formula, and the two Excel figures.

Open

Notebook render

Notebook

R Notebook HTML of the same walkthrough, with the source embed and the workbook screenshots inline.

Open

R Markdown source

R

Knit this file to rebuild the HTML and overwrite the sample workbook.

Source

R GLM in Excel

Excel

Workbook with the coefficients sheet and the R predictions, ready for the verification formula.

Download