modelr

modelr is an R package that provides functions for creating elegant modeling pipelines with tidyverse-style syntax. It was designed to teach modeling basics and is now superseded by the tidymodels framework.

The package offers resampling methods (bootstrap, k-fold cross-validation, Monte Carlo cross-validation), model quality metrics (RMSE, R-squared, MAE), and utilities for adding predictions and residuals to data frames. It simplifies common modeling workflows like creating training/test splits, generating evenly-spaced prediction grids, and evaluating model performance. These tools work seamlessly with pipes and tidy data principles, making it straightforward to integrate modeling steps into data analysis workflows.