Getting Started with R for Data Science in 2026: Installation, Setup, and Your First Analysis

From zero to your first working analysis — install R, set up RStudio, and learn the core tools every data scientist needs.

Last Updated 2026 — originally published July 2018

---

R is still one of the most powerful tools in a data scientist’s toolkit — especially for statistical modeling, EDA, and clinical research. This guide covers everything from installation to your first working analysis, updated for 2025.

Python gets more headlines, but R remains indispensable for several categories of work: statistical modeling, clinical trial analysis, bioinformatics, and exploratory data visualization. Many clinical researchers and biostatisticians work primarily in R, which means knowing R makes you a better collaborator in medical AI projects.

This guide is for anyone starting with R from scratch, or refreshing their knowledge after a gap.

---

Table of Contents

• Should You Learn R or Python? (Honest Answer)
• Installing R and RStudio
• Essential Packages to Install First
• Understanding R’s Core Data Structures
• Your First Real Analysis: Loading, Exploring, Cleaning
• Data Manipulation with dplyr
• Data Visualization with ggplot2
• Connecting R to Python Workflows
• Learning Resources

---

Should You Learn R or Python? (Honest Answer)

Learn both, but prioritize based on your immediate needs:

Task	Better choice
Deep learning / neural networks	Python (PyTorch, TensorFlow)
Quick statistical modeling (linear regression, mixed models)	R
Exploratory data visualization	R (ggplot2 is superior)
Clinical trial analysis	R (gold standard in pharma)
Bioinformatics (RNA-seq, GWAS)	R (Bioconductor ecosystem)
Production ML systems	Python
Rapid prototyping of ML models	Python

For researchers at the intersection of AI and medicine: learn Python for model building, and R for statistical analysis and clinical collaboration. I use both in my own research.

---

Installing R and RStudio

Step 1: Install R

Windows and Mac:
Download from https://cran.r-project.org/ — choose your OS, download the installer, run it.

Ubuntu/Linux:

sudo apt update
sudo apt install r-base r-base-dev

For a more current version on Ubuntu:

# Add CRAN repository for latest R
wget -qO- https://cloud.r-project.org/bin/linux/ubuntu/marutter_pubkey.asc | sudo gpg --dearmor -o /usr/share/keyrings/r-project.gpg
echo "deb [signed-by=/usr/share/keyrings/r-project.gpg] https://cloud.r-project.org/bin/linux/ubuntu $(lsb_release -cs)-cran40/" | sudo tee /etc/apt/sources.list.d/r-project.list
sudo apt update
sudo apt install r-base

Step 2: Install RStudio

RStudio is the IDE for R. Download from: https://posit.co/download/rstudio-desktop/

RStudio provides:

• Script editor with syntax highlighting and autocomplete
• Interactive console
• Integrated plots panel
• Environment viewer (see all variables and their values)
• Help panel

---

Essential Packages to Install First

Open RStudio and run this in the console:

# The tidyverse: a collection of packages that work together
install.packages("tidyverse")  # includes dplyr, ggplot2, tidyr, readr, and more

# Statistical modeling
install.packages("caret")       # ML training and evaluation
install.packages("glmnet")      # Regularized regression
install.packages("survival")    # Survival analysis (essential for clinical data)

# Handling imbalanced data
install.packages("ROSE")

# Visualization extras
install.packages("corrplot")    # Correlation heatmaps
install.packages("ggpubr")      # Publication-ready plots

# Working with Excel/CSV
install.packages("readxl")
install.packages("openxlsx")

---

Understanding R’s Core Data Structures

Unlike Python, R was designed for statistical computing, so its data structures reflect that:

# Vectors — the atomic unit (like Python lists but typed)
nums <- c(1, 2, 3, 4, 5)
chars <- c("hello", "world")
logical <- c(TRUE, FALSE, TRUE)

# Data frames — the equivalent of pandas DataFrames
df <- data.frame(
  name = c("Alice", "Bob", "Carol"),
  age = c(28, 35, 22),
  score = c(0.87, 0.91, 0.79)
)

# Tibbles — modern, improved data frames (from tidyverse)
library(tibble)
tb <- tibble(name = c("Alice", "Bob"), age = c(28, 35))

# Factors — categorical variables with defined levels
treatment <- factor(c("control", "treatment", "control"),
                    levels = c("control", "treatment"))

The most important thing to understand: R is vectorized. Most operations apply to entire vectors without explicit loops:

nums * 2          # [1] 2 4 6 8 10 — no for loop needed
nums > 3          # [1] FALSE FALSE FALSE TRUE TRUE
sum(nums > 3)     # [1] 2 — count of values greater than 3

---

Your First Real Analysis

Let’s work through a real example using the built-in mtcars dataset (car performance data):

# Load data
data(mtcars)

# Step 1: Inspect
dim(mtcars)          # 32 rows, 11 columns
str(mtcars)          # data types
summary(mtcars)      # min, max, mean, quartiles
head(mtcars, 6)      # first 6 rows

# Step 2: Check missing values
colSums(is.na(mtcars))  # none in this dataset

# Step 3: Distribution of key variable
hist(mtcars$mpg,
     main = "Distribution of Miles Per Gallon",
     xlab = "MPG", col = "steelblue", breaks = 10)

# Step 4: Correlation
cor(mtcars$mpg, mtcars$wt)  # negative: heavier cars get fewer mpg

# Step 5: Simple linear model
model <- lm(mpg ~ wt + hp + cyl, data = mtcars)
summary(model)
# Shows coefficients, p-values, R-squared

---

Data Manipulation with dplyr

dplyr is the R equivalent of pandas for data manipulation. Its “verb” syntax is clean and composable:

library(dplyr)

# The pipe operator %>% chains operations
mtcars %>%
  filter(cyl == 6) %>%            # keep only 6-cylinder cars
  select(mpg, wt, hp) %>%         # keep only these columns
  mutate(wt_kg = wt * 453.6) %>%  # add new column
  arrange(desc(mpg)) %>%          # sort by mpg descending
  head(5)                          # first 5 rows

# Group and summarize
mtcars %>%
  group_by(cyl) %>%
  summarise(
    avg_mpg = mean(mpg),
    avg_hp = mean(hp),
    count = n()
  )

# Join two data frames (like SQL JOIN)
left_join(df1, df2, by = "id")

---

Data Visualization with ggplot2

ggplot2 is built on the “grammar of graphics” — you build plots layer by layer:

library(ggplot2)

# Scatter plot
ggplot(mtcars, aes(x = wt, y = mpg, color = factor(cyl))) +
  geom_point(size = 3, alpha = 0.7) +
  geom_smooth(method = "lm", se = TRUE) +
  labs(title = "Car Weight vs Fuel Efficiency",
       x = "Weight (1000 lbs)", y = "Miles per Gallon",
       color = "Cylinders") +
  theme_minimal()

# Box plot comparing groups
ggplot(mtcars, aes(x = factor(cyl), y = mpg, fill = factor(cyl))) +
  geom_boxplot(alpha = 0.7, outlier.color = "red") +
  labs(title = "MPG by Number of Cylinders",
       x = "Cylinders", y = "MPG") +
  theme_minimal() +
  theme(legend.position = "none")

# Histogram with density
ggplot(mtcars, aes(x = mpg)) +
  geom_histogram(aes(y = ..density..), bins = 15,
                 fill = "steelblue", alpha = 0.7) +
  geom_density(color = "darkred", size = 1.2) +
  labs(title = "MPG Distribution") +
  theme_minimal()

---

Connecting R to Python Workflows

If your modeling happens in Python but your statistics or visualization happens in R, you have two good options:

Option 1: Export/import via CSV

# Export from R
write.csv(df_cleaned, "data_for_python.csv", row.names = FALSE)

# Import in Python
import pandas as pd
df = pd.read_csv("data_for_python.csv")

Option 2: Use reticulate (R package that calls Python)

install.packages("reticulate")
library(reticulate)
py_run_string("import pandas as pd; df = pd.read_csv('data.csv')")

Option 3: RMarkdown / Quarto
If you write research reports, Quarto documents can contain both R and Python code chunks in the same document — ideal for reproducible research.

---

Learning Resources

• R for Data Science (r4ds.had.co.nz) — free online book by Hadley Wickham; the definitive tidyverse guide
• MIT Analytics Edge on edX — the course that got me started; uses R for real business analytics problems
• R-bloggers (r-bloggers.com) — daily community tutorials
• Swirl — learn R interactively inside R itself: install.packages("swirl"); swirl()
• Tidytuesday — weekly social data project using R; great for building visualization skills

---