R Programming Tutorial - Learn the Basics of Statistical Computing

Key Takeaways

• R's Prominence: R is the most used software by data mining experts, surpassing Python by 50%, primarily because it's free, open-source, optimized for vector operations (reducing the need for explicit loops), and supported by a vast community and over 9,000 contributed packages. 0:33
• Setting Up Your Environment: Install R by downloading from r-project.org (using the "cloud" mirror for convenience), then install RStudio from rstudio.com (desktop free version) which provides a unified, more organized interface over the basic R application. 2:07
• Understanding R's Interface: The R application (or RStudio's script window) allows you to write and run code (Command/Control + Enter), while the console displays output; lines starting with # are comments. 4:21
• Managing Packages: R packages extend functionality, with base packages included by default and contributed packages needing to be downloaded, installed, and loaded; use CRAN (cran.r-project.org) or CRANtastic (crantastic.org) for discovery, or a package manager like Pacman (pacman::p_load()) to streamline installation and loading of multiple packages (e.g., dplyr, ggplot2, rio, rmarkdown). 11:52
• Basic Data Visualization: The plot() command is versatile, adapting to the data types and number of variables for quick visualizations (e.g., bar charts for categorical, scatterplots for quantitative pairs, box plots for categorical vs. quantitative, or a matrix of plots for entire datasets), and can be customized with arguments like col, pch, main, xlab, ylab for publication-quality output. 19:09
• Specific Chart Types: Use hist() for quantitative variables to examine distribution shape, gaps, and outliers; barplot() for categorical data requires a pre-calculated summary table (e.g., using table()) from raw data. 29:52
• Accessing and Manipulating Data: Data exists in different types (numeric, character, logical) and structures (vector, matrix, array, data frame, list); understand coercion (changing data types/structures) and factors (attributing possible values and order to categorical vectors); manually enter data using operators like : (sequence), c() (concatenate), scan() (live input), or rep() (repetition), and import files (CSV, TXT, Excel) easily with the rio package's import() function. 1:06:14
• Fundamental Statistical Analysis: The summary() function provides quick descriptive statistics (frequencies for categorical, quartiles/mean for quantitative); for more detailed metrics (e.g., standard deviation, skewness, kurtosis), use the describe() function from the psych package. 52:21
• Selecting and Subsetting Data: Focus analysis by selecting cases based on categories (e.g., iris[iris$species == "setosa",]), numerical values (e.g., iris[iris$petal.length < 2,]), or multiple criteria using &, and save these selections into new data frames for repeated use. 1:00:18
• Introduction to Data Modeling: R supports various modeling techniques like hierarchical clustering (grouping similar cases), principal component analysis (PCA for dimensionality reduction and variable grouping), and regression (predicting an outcome variable from multiple predictors with functions like lm()). 1:42:30