R Introduction Lecture Notes

Course Overview

• Instructor: Barton Poulsen
• Goal: Introduce R and its applications in data science.
• Importance of R:
  • Ranked first in a survey of data mining experts for data science software, outperforming Python by 50%.

Reasons to Learn R

• Cost-effective: Free and open-source, unlike many alternatives that cost thousands annually.
• Optimized for Vector Operations: Allows processing entire rows/tables without explicit loops.
• Strong Community Support: Access to examples, assistance, and continuous development.
• Extensive Package Availability: Over 9,000 contributed packages enhance R's functionality.

Getting Started with R

Installing R

1. Go to r-project.org.
2. Click on "Download R" to access CRAN.
3. Choose and download the appropriate installer for your operating system (Mac, Windows, Linux).
4. Install R by following standard installation procedures.

Course Files

• Download course files from the provided link and unzip them.
• Place the folder containing .R scripts and data files on your desktop.

Using R Environment

Basic R Interface

• Script Window: Where you write your code.
• Console Window: Displays output and results.
• Commented Lines: Begin with # and are not executed.

Running Code Example

• Load datasets and explore functionality using the well-known Iris dataset:
  • View first six lines using head(iris).
  • Get summary statistics with summary(iris).
  • Create basic plots with plot() commands.

Introduction to RStudio

• RStudio: A user-friendly interface that organizes R's functionality better than the basic R interface.
• Setting up RStudio: Download from rstudio.com and install.
• RStudio Interface: Consists of script, console, environment, and plots panes for easy navigation and management.

Working with R Packages

Importance of Packages

• Packages are collections of code that extend R's capabilities.
• Two types of packages:
  • Base Packages: Installed with R but not loaded by default.
  • Contributed Packages: Need to be downloaded and loaded separately.

Finding and Installing Packages

• Use CRAN, CRANTASTIC, or GitHub to find packages.
• Commonly used packages include:
  • dplyr: Data manipulation.
  • tidyr: Data cleaning.
  • ggplot2: Data visualization.

Loading Packages in R

• Use library(package_name) to load packages.
• To install a package, use install.packages("package_name").

Basic Graphics in R

Creating Basic Plots

• Use plot() for scatterplots of quantitative data.
• Create bar charts for categorical data using barplot().
• Use histograms for quantitative data with hist() commands.

Visualizing Data

• Create multiple graphs for comparative analysis (small multiples).
• Overlaid plots can provide richer data visualization.
• Ensure clarity and avoid clutter in graphs.

Basic Statistics in R

Summary Statistics

• Use summary() to get basic statistics for datasets.
• describe() from the psych package for detailed statistics.

Selecting Cases

• Filter datasets by categories or values using conditional indexing.

Importing Data in R

Data Formats

• R can import various formats:
  • CSV, TXT, XLSX, JSON.
• Recommended to use the Rio package for ease of import.

Importing Example

1. Use import("file_path") for CSV/TXT.
2. Use import("file_path.xlsx") for Excel files.
3. Explore data using View(data_frame).

Modeling Data in R

Hierarchical Clustering

• Group data points based on similarity.
• Use functions like dist() and hclust() for clustering.

Principal Component Analysis (PCA)

• Reduces dimensionality by transforming to fewer components.
• Implemented using prcomp() function.

Regression Analysis

• Predict an outcome variable using multiple predictor variables.
• Use lm() to create linear models.
• Explore different regression techniques (stepwise, lasso, etc.).

Conclusion and Next Steps

• Explore additional resources and courses at datalab.cc.
• Consider comparing R with Python for broader programming skills.
• Engage with visualization concepts to enhance data presentation.
• Explore machine learning methods for advanced data processing.