pkg.drugage

The goal of pkg.drugage is to provide functions to streamline the creation of data classification models. This package was originally developed with the intention of creating a classificiation model able to categorize data into youth vs. adult groups based on the trend of drug use/frequency, but its functions can be applied to any similar scenario.

Table of Contents

• Ecosystem Context
• Functions in the Package
• Dependencies
• Installation
• Example Use
• Code of Conduct
• Licenses

Ecosystem Context

In the R ecosystem, pkg.drugage complements existing packages such as:

• tidymodels: While tidymodels offers a comprehensive framework for modeling, pkg.drugage provides convenience functions specifically tailored for classification workflows, including pre-processing steps and visualization helpers.
• caret: Similar to caret, pkg.drugage supports training and evaluating models, but focuses more narrowly on binary classification use cases and integrates more tightly with age-group classification logic.
• janitor and skimr: These offer data cleaning and summary tools, while pkg.drugage includes domain-specific cleaning for age-group classification in addition to generalized preprocessing.

Unlike broader frameworks, pkg.drugage prioritizes accessibility, domain focus, and end-to-end reproducibility for education and research scenarios.

Functions

• download_data: Downloads a file from a given URL and saves it to a specified destination.
• create_directory: Checks if the specified directory exists, and creates the directory path if necessary.
• data_cleaning: Cleans the dataset, classifies individuals into age groups and saves the cleaned dataset to a given path.
• eda_functions: Performs exploratory data analysis (EDA) on the cleaned dataset, such as creating bar and scatter plots to visualize the raw data.
• train_and_predict: Takes a model specification and a recipe, fits the model on the training data, and returns class predictions on the testing data.
• create_confusion_matrix: Takes model predictions, generates a confusion matrix and saves it as a PNG figure and as a CSV table in the specified output directory. Returns the ggplot object used to create the plot.

Dependencies

R Dependencies

• docopt: 0.7.1
• dplyr: 1.1.4
• parsnip: 1.3.1
• pointblank: 0.12.2
• readr: 2.1.5
• recipes: 1.1.1
• rsample: 1.2.1
• tidyr: 1.3.1
• workflows: 1.2.0

These dependencies will be automatically installed when running pak::pak("DSCI-310-2025/pkg.drugage") to install the package.

Installation

You can install the development version of pkg.drugage from GitHub with:

install.packages("pak")

pak::pak("DSCI-310-2025/pkg.drugage")

Then load the package:

library(pkg.drugage)

Example

The following is a basic example which shows you how to use some of the functions:

library(pkg.drugage)

# create sample data
sample_data <- data.frame(
 age = c("18-24", "25-34", "35-44"),
 alcohol.use = c(80, 75, 60),
 alcohol.frequency = c(20, 15, 10),
 marijuana.use = c(50, 30, 20),
 heroin.frequency = c(5, 3, 2),
 class = c("youth", "adult", "adult"),
 n = c(100, 200, 150))
 
# Clean data
cleaned_data <- clean_drug_use_data(sample_data)
classified_data <- classify_age_group(cleaned_data)
head(classified_data)
#>     age alcohol.use alcohol.frequency marijuana.use heroin.frequency class   n
#> 1 18-24          80                20            50                5 adult 100
#> 2 25-34          75                15            30                3 adult 200
#> 3 35-44          60                10            20                2 adult 150

# EDA
create_bar_use_plot(
sample_data,
"age",
"alcohol.use",
"Alcohol Use",
"Age",
"Proportion",
"dodgerblue",
"man/figures/README-test1.png")

sample_aggregated <- aggregate_data(sample_data)

create_grouped_bar_plot(
sample_aggregated,
"Youth vs. Adult Comparison",
"Substance Type",
"Mean Substance Use (%)",
"man/figures/README-test6.png"
)

Code of Conduct

Please note that pkg.drugage is released with a Code of Conduct. By contributing to this project, you agree to abide by its terms.

Licenses

This project is dual-licensed under the following:

• Creative Commons Zero v1.0 Universal (CC0 1.0) You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission.

• MIT License
  A short and permissive software license. You can use it in proprietary software provided that all copies include the license terms and the copyright notice.