Getting Started with gooseR

In gooseR: R Integration for 'Goose' AI

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  eval = FALSE
)

library(gooseR)

Welcome to gooseR! 🦆

gooseR brings the power of AI directly into your R workflow. Whether you're analyzing data, writing code, or creating visualizations, gooseR is your intelligent assistant that learns from your preferences and helps you work more efficiently.

Installation and Setup

Installing gooseR

# Install from GitHub (CRAN submission pending)
# install.packages("remotes")
remotes::install_github("blockbtheriault/gooseR")

Testing Your Setup

The first step is to ensure gooseR can communicate with the goose CLI:

library(gooseR)

# Test the connection
if (goose_test_cli()) {
  message("✅ Goose CLI is ready!")
} else {
  message("❌ Need to configure goose - see next section")
}

Configuration (if needed)

If you don't have goose CLI already configured, you'll need to set up your AI provider:

# For OpenAI users
goose_configure(
  provider = "openai",
  model = "gpt-4o",
  api_key = "your-api-key-here"
)

# For Anthropic Claude users
goose_configure(
  provider = "anthropic",
  model = "claude-3-opus",
  api_key = "your-api-key-here"
)

Your First gooseR Session

1. Ask Questions

The simplest way to start is by asking goose questions:

# Ask about your data
goose_ask("What are the key characteristics of the mtcars dataset?")

# Get analysis suggestions
goose_ask("What statistical tests would be appropriate for comparing mpg across different numbers of cylinders?")

# Request code examples
response <- goose_ask("Show me how to create a correlation matrix heatmap in R")
cat(response)

2. Get Code Reviews

One of gooseR's most powerful features is intelligent code review. Unlike generic linters, goose actually reads and understands your code:

# Write some code
my_analysis <- function(data) {
  # Calculate mean without checking for NA
  avg <- mean(data$value)

  # Using a loop instead of vectorized operation
  results <- c()
  for(i in 1:nrow(data)) {
    results[i] <- data$value[i] * 2
  }

  return(list(avg = avg, doubled = results))
}

# Get a gentle review
goose_honk(severity = "gentle")

# Get more critical feedback
goose_honk(severity = "moderate")

# For tough love
goose_honk(severity = "harsh")

3. Save Your Work with Memory

gooseR's memory system lets you save any R object and retrieve it later - even across sessions:

# Create a model
model <- lm(mpg ~ wt + cyl + hp, data = mtcars)

# Save it with tags for easy retrieval
goose_save(
  model,
  category = "models",
  tags = c("mtcars", "regression", "fuel_efficiency")
)

# List saved objects
goose_list(category = "models")

# Load it back (even in a new session)
my_model <- goose_load("model")
summary(my_model)

4. Create Beautiful Visualizations

Apply professional branding to your plots instantly:

library(ggplot2)

# Create a plot with Block branding
p <- ggplot(mtcars, aes(x = wt, y = mpg)) +
  geom_point(size = 3, alpha = 0.7) +
  geom_smooth(method = "lm", se = TRUE) +
  theme_brand("block") +  # Apply Block theme
  labs(
    title = "Fuel Efficiency vs Weight",
    subtitle = "Linear relationship in mtcars dataset",
    x = "Weight (1000 lbs)",
    y = "Miles per Gallon"
  )

print(p)

# Access brand colors
colors <- brand_palette("block", "categorical")

Essential Workflows

For Data Analysis

# Load your data
my_data <- read.csv("my_dataset.csv")

# Share a sample with goose for context
goose_give_sample(my_data)

# Get an analysis plan
plan <- goose_make_a_plan("exploratory")
cat(plan)

# Do your analysis...
# ...

# Get feedback on your approach
goose_honk(severity = "moderate")

# Save your work for tomorrow
goose_continuation_prompt()

For Survey Data

# Load survey data with long question names
survey <- read.csv("qualtrics_export.csv")

# Automatically rename columns intelligently
clean_survey <- goose_rename_columns(survey)

# View the mapping
goose_view_column_map(clean_survey)
# "How satisfied are you with our customer service?" → "sat_cust_serv"
# "On a scale of 1-10, how likely are you to recommend..." → "nps"

For Team Collaboration

# Before starting work, backup existing objects
goose_backup()

# Work in a temporary session that auto-cleans
with_goose_session({
  # Experimental work here
  test_model <- lm(mpg ~ ., data = mtcars)
  goose_save(test_model, category = "temp", tags = "experiment")

  # This will be auto-cleaned when session ends
}, cleanup = TRUE)

# Create a handoff document for your colleague
goose_handoff()

# Clean up test objects
goose_clear_tags(c("test", "temp", "draft"))

Tips for Success

1. Start Simple: Begin with goose_ask() to get comfortable with AI responses

2. Use Severity Levels: Start with "gentle" code reviews and work up to "harsh" as you get comfortable

3. Tag Everything: Use descriptive tags when saving objects - you'll thank yourself later

4. Share Context: Use goose_give_sample() before asking for analysis help

5. Save Your Work: Use goose_continuation_prompt() at the end of sessions

Next Steps

• Read the Survey Data Analysis vignette for survey-specific workflows
• See Code Review and Testing for development best practices
• Explore the Memory Workflow vignette for advanced memory management

Getting Help

# Get help on any function
?goose_ask
?goose_honk
?goose_save

# Ask goose for help!
goose_ask("How do I use goose_rename_columns with custom abbreviations?")

# Check your gooseR version
packageVersion("gooseR")

Welcome to the gooseR community! 🦆

gooseR documentation built on Feb. 6, 2026, 5:07 p.m.