Welcome!

Tidymodels is a collection of packages developed for modeling and machine learning. These packages simplify the task of fitting a model to data by providing tools to assist with data pre-proprocessing, model initiation, and results valiation. Tidymodels emphasizes simple, readable code by abstracting complex modelling operations into common interfaces and optimizing functions for the pipe operator.

To use Tidymodels, simply install and load the tidymodels package. The tidymodels package loads all of its core modeling packages as well as some helpful tidyverse packages such as dplyr and ggplot2. Find a complete list of Tidymodels packages here.

library(tidymodels)

Overview

Each tab on this site corresponds to a different package from Tidymodels:

• RSample - Resample and split training/testing data
• Recipes - Create “pre-processing objects” that apply transformations to data
• Parsnip - Common, simple interface for model initiation
• Yardstick - Evaluate model performance
• Workflows - Bundle recipes and models into a single object
• Tune - Tune hyperparameters for models
• Workflowsets - Create and compare sets of workflows

These pages are meant to be read in order, as modeling examples are built cumulatively. To view abbreviated versions of the code through each tutorial, view the following tabs:

• Iris - Predicting Iris species from their measurements.
• Cfin - Optimizing and comparing models of C. Finmarchicus abundance.

General Resources

• https://rviews.rstudio.com/2019/06/19/a-gentle-intro-to-tidymodels/
  • A simple walkthrough of some core Tidymodels packages and concepts.
• https://www.tmwr.org/
  • A highly detailed guidebook of Tidymodels and how to create accurate models. Much of this tutorial site is just a simplifed rehash of information from this guidebook – if you want to take a deep dive into tuning and optimizing your model, I highly recommend going through this resource.

About

This site is authored by Omi Johnson and is meant to serve as an introductory walkthrough of Tidymodels. This site was built using RMarkdown and GitHub hosting – find a simple walkthrough of how to build your own RMarkdown site here. The source repository for this site is on my GitHub.

Questions? Comments? Contact me by email at ojohnson@bigelow.org

NOTE: This is not a comprehensive guide. These notes are incomplete and only mention methods/techniques that the author is familiar with. Descriptions of packages and methods most likely do not include all possible arguments or applications. For a complete list of each package’s methods and capabilities, please look at the official documentation.