2 Data Basics

2.1 Introduction

This chapter covers the fundamental concepts of working with data in R. You’ll learn how to import, clean, and prepare data for analysis, which are essential skills for any data analysis project across all natural science disciplines.

2.2 Understanding Data Structures

Before diving into data analysis, it’s important to understand the basic data structures in R:

2.2.1 Data Types

R has several basic data types:

Numeric: Decimal values (e.g., measurements of temperature, pH, concentration, or distance)
Integer: Whole numbers (e.g., counts of organisms, samples, or observations)
Character: Text strings (e.g., species names, site descriptions, or treatment labels)
Logical: TRUE/FALSE values (e.g., presence/absence data or condition met/not met)
Factor: Categorical variables with levels (e.g., experimental treatments, taxonomic classifications, or soil types)
Date/Time: Temporal data (e.g., sampling dates, observation times, or seasonal markers)

Code

# Examples of different data types
numeric_example <- 25.4  # Temperature in Celsius
character_example <- "Adelie"  # Penguin species
logical_example <- TRUE  # Presence/absence data
factor_example <- factor(c("Control", "Treatment", "Control"),
                         levels = c("Control", "Treatment"))
date_example <- as.Date("2020-07-15")  # Sampling date

# Print examples
print(numeric_example)

[1] 25.4

Code

print(character_example)

[1] "Adelie"

Code

print(logical_example)

[1] TRUE

Code

print(factor_example)

[1] Control   Treatment Control  
Levels: Control Treatment

Code

print(date_example)

[1] "2020-07-15"

Code Explanation

This code demonstrates the fundamental data types in R:

Numeric Data:
- Created a decimal value representing temperature in Celsius
- Common for environmental measurements and continuous data
Character Data:
- Created a text string for a penguin species name
- Used for categorical variables like species names or site descriptions
Logical Data:
- Created a TRUE value representing presence/absence
- Used for binary data or conditions in analyses
Factor Data:
- Created an ordered categorical variable with two levels
- Explicitly defined factor levels (“Control” and “Treatment”)
- Essential for statistical analyses and proper plotting order
Date Data:
- Created a date object using the as.Date() function
- Used for temporal data in ecological studies

Results Interpretation

The output shows how R stores and displays different data types:

Numeric Data:
- Displayed as 25.4 without type indication
- R treats this as a continuous numeric value for calculations
Character Data:
- Displayed as “Adelie” with quotation marks indicating text
- Cannot be used for numerical operations
Logical Data:
- Displayed as TRUE (without quotation marks)
- Can be used in conditional operations and converts to 1 (TRUE) or 0 (FALSE) in calculations
Factor Data:
- Displayed with levels information: Control, Treatment, Control
- Internally stored as integers with labels
- Order of levels is preserved as specified
Date Data:
- Displayed in standardized YYYY-MM-DD format
- Allows for time-based calculations and comparisons

PROFESSIONAL TIP: Data Management Best Practices

Proper data management is critical for reproducible research in natural sciences:

Document metadata: Always maintain detailed records about data collection methods, units, and variable definitions
Use consistent naming conventions: Create clear, consistent file and variable names (e.g., site_01_temp_2023.csv instead of data1.csv)
Preserve raw data: Never modify your original data files; always work with copies for cleaning and analysis
Version control: Use Git or similar tools to track changes to your data processing scripts
Implement quality control: Create automated checks for impossible values, outliers, and inconsistencies
Plan for missing data: Develop a consistent strategy for handling missing values before analysis begins
Create tidy data: Structure data with one observation per row and one variable per column
Use open formats: Store data in non-proprietary formats (CSV, TSV) for long-term accessibility
Back up regularly: Maintain multiple copies of your data in different physical locations
Consider data repositories: Share your data through repositories like Dryad, Zenodo, or discipline-specific databases

2.2.2 Data Structures in R

R has several data structures for organizing information:

Code

# Load real datasets
library(readr)
penguins <- read_csv("../data/environmental/climate_data.csv")
crops <- read_csv("../data/agriculture/crop_yields.csv")

# Vector example - penguin bill lengths
bill_lengths <- na.omit(penguins$bill_length_mm[1:10])
print(bill_lengths)

[1] 39.1 39.5 40.3 36.7 39.3 38.9 39.2 34.1 42.0
attr(,"na.action")
[1] 4
attr(,"class")
[1] "omit"

Code

# Matrix example - create a matrix from penguin measurements
penguin_matrix <- as.matrix(penguins[1:5, 3:6])
print(penguin_matrix)

     bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
[1,]           39.1          18.7               181        3750
[2,]           39.5          17.4               186        3800
[3,]           40.3          18.0               195        3250
[4,]             NA            NA                NA          NA
[5,]           36.7          19.3               193        3450

Code

# Data frame example - first few rows of penguin data
penguin_data <- penguins[1:5, ]
print(penguin_data)

# A tibble: 5 × 8
  species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
  <chr>   <chr>              <dbl>         <dbl>             <dbl>       <dbl>
1 Adelie  Torgersen           39.1          18.7               181        3750
2 Adelie  Torgersen           39.5          17.4               186        3800
3 Adelie  Torgersen           40.3          18                 195        3250
4 Adelie  Torgersen           NA            NA                  NA          NA
5 Adelie  Torgersen           36.7          19.3               193        3450
# ℹ 2 more variables: sex <chr>, year <dbl>

Code

# List example - store different aspects of the dataset
penguin_summary <- list(
  species = unique(penguins$species),
  avg_bill_length = mean(penguins$bill_length_mm, na.rm = TRUE),
  sample_size = nrow(penguins),
  years = unique(penguins$year)
)
print(penguin_summary)

$species
[1] "Adelie"    "Gentoo"    "Chinstrap"

$avg_bill_length
[1] 43.92193

$sample_size
[1] 344

$years
[1] 2007 2008 2009

Code Explanation

This code demonstrates the main data structures in R using real ecological datasets:

Data Loading:
- Uses readr::read_csv() to import real datasets on penguins and crop yields
- Loads data with proper data types and handling
Vector Creation:
- Creates a numeric vector of bill lengths
- Uses na.omit() to remove missing values
- Subsets only the first 10 values with [1:10]
Matrix Construction:
- Creates a numeric matrix from penguin measurements
- Uses as.matrix() to convert data frame columns to matrix
- Selects rows 1-5 and columns 3-6 using indexing
Data Frame Handling:
- Demonstrates a data frame (the most common data structure)
- Shows how to subset rows while keeping all columns
List Creation:
- Creates a list to store heterogeneous data elements
- Contains different data types: character vector, numeric value, and integer
- Demonstrates how lists can store complex, nested information

Results Interpretation

The output reveals the structure and properties of different R data types:

Vector Output:
- Shows a one-dimensional array of bill length measurements
- All elements are of the same type (numeric)
- Suitable for storing a single variable’s values
Matrix Output:
- Displays a two-dimensional array of measurements
- All values must be of the same type (converted to numeric)
- Row and column indices are shown
- Efficient for mathematical operations but less flexible than data frames
Data Frame Output:
- Shows a tabular structure with different variable types
- Preserves column names and data types
- The foundation of most data analysis in R
- Each column can have a different data type
List Output:
- Displays a collection of disparate elements
- Shows the flexibility of lists for storing mixed data
- Demonstrates named elements for easy access
- Ideal for storing complex results and heterogeneous data

2.3 Importing Data

2.3.1 Reading Data Files

R provides several functions for importing data from different file formats:

Code

# CSV files - Palmer Penguins dataset
penguins_csv <- read.csv("../data/environmental/climate_data.csv")
head(penguins_csv, 3)

  species    island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
1  Adelie Torgersen           39.1          18.7               181        3750
2  Adelie Torgersen           39.5          17.4               186        3800
3  Adelie Torgersen           40.3          18.0               195        3250
     sex year
1   male 2007
2 female 2007
3 female 2007

Code

# Using the tidyverse approach for better handling
library(tidyverse)
penguins_tidy <- readr::read_csv("../data/environmental/climate_data.csv")
head(penguins_tidy, 3)

# A tibble: 3 × 8
  species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
  <chr>   <chr>              <dbl>         <dbl>             <dbl>       <dbl>
1 Adelie  Torgersen           39.1          18.7               181        3750
2 Adelie  Torgersen           39.5          17.4               186        3800
3 Adelie  Torgersen           40.3          18                 195        3250
# ℹ 2 more variables: sex <chr>, year <dbl>

Code

# Crop yields dataset
crops_csv <- read.csv("../data/agriculture/crop_yields.csv")
head(crops_csv, 3)

       Entity Code Year Wheat..tonnes.per.hectare. Rice..tonnes.per.hectare.
1 Afghanistan  AFG 1961                     1.0220                     1.519
2 Afghanistan  AFG 1962                     0.9735                     1.519
3 Afghanistan  AFG 1963                     0.8317                     1.519
  Maize..tonnes.per.hectare. Soybeans..tonnes.per.hectare.
1                      1.400                            NA
2                      1.400                            NA
3                      1.426                            NA
  Potatoes..tonnes.per.hectare. Beans..tonnes.per.hectare.
1                        8.6667                         NA
2                        7.6667                         NA
3                        8.1333                         NA
  Peas..tonnes.per.hectare. Cassava..tonnes.per.hectare.
1                        NA                           NA
2                        NA                           NA
3                        NA                           NA
  Barley..tonnes.per.hectare. Cocoa.beans..tonnes.per.hectare.
1                        1.08                               NA
2                        1.08                               NA
3                        1.08                               NA
  Bananas..tonnes.per.hectare.
1                           NA
2                           NA
3                           NA

Code Explanation

This code demonstrates different methods for importing data in R:

Base R Import:
- Uses read.csv() from base R to import the penguin dataset
- Simple approach that works without additional packages
- Generally slower for large datasets and less flexible with column types
Tidyverse Import:
- Uses readr::read_csv() from the tidyverse ecosystem
- More efficient for large datasets and better type inference
- Maintains consistent column types and handles problematic values better
Data Preview:
- Uses head() with argument 3 to display just the first three rows
- Allows quick inspection of data structure without overwhelming output
- Essential first step to verify successful import and correct structure
Multiple Datasets:
- Demonstrates importing different datasets (penguins and crop yields)
- Shows the same approach works across various data sources

Results Interpretation

The output reveals differences between import methods and gives insight into the datasets:

Data Structure Visibility:
- Both datasets show proper column names and values
- The tidyverse import (readr) provides cleaner output with column types
- Types are indicated ( for numeric, for character, etc.)
Import Method Comparison:
- Base R (read.csv()) and tidyverse (read_csv()) produce similar results
- Tidyverse version provides more metadata about column types
- Both successfully imported the data with proper structure
Data Content Preview:
- Penguin data contains morphological measurements and categorical variables
- Crop yield data includes countries, years, and production statistics
- Both datasets appear properly formatted for analysis

2.3.2 Exploring Real-World Datasets

Let’s explore some of the real-world datasets we have available:

Code

# Palmer Penguins dataset
penguins <- read_csv("../data/environmental/climate_data.csv")
glimpse(penguins)

Rows: 344
Columns: 8
$ species           <chr> "Adelie", "Adelie", "Adelie", "Adelie", "Adelie", "A…
$ island            <chr> "Torgersen", "Torgersen", "Torgersen", "Torgersen", …
$ bill_length_mm    <dbl> 39.1, 39.5, 40.3, NA, 36.7, 39.3, 38.9, 39.2, 34.1, …
$ bill_depth_mm     <dbl> 18.7, 17.4, 18.0, NA, 19.3, 20.6, 17.8, 19.6, 18.1, …
$ flipper_length_mm <dbl> 181, 186, 195, NA, 193, 190, 181, 195, 193, 190, 186…
$ body_mass_g       <dbl> 3750, 3800, 3250, NA, 3450, 3650, 3625, 4675, 3475, …
$ sex               <chr> "male", "female", "female", NA, "female", "male", "f…
$ year              <dbl> 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007…

Code

# Basic summary statistics
summary(penguins$bill_length_mm)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  32.10   39.23   44.45   43.92   48.50   59.60       2

Code

summary(penguins$flipper_length_mm)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  172.0   190.0   197.0   200.9   213.0   231.0       2

Code

# Crop yields dataset
crops <- read_csv("../data/agriculture/crop_yields.csv")
glimpse(crops)

Rows: 13,075
Columns: 14
$ Entity                             <chr> "Afghanistan", "Afghanistan", "Afgh…
$ Code                               <chr> "AFG", "AFG", "AFG", "AFG", "AFG", …
$ Year                               <dbl> 1961, 1962, 1963, 1964, 1965, 1966,…
$ `Wheat (tonnes per hectare)`       <dbl> 1.0220, 0.9735, 0.8317, 0.9510, 0.9…
$ `Rice (tonnes per hectare)`        <dbl> 1.5190, 1.5190, 1.5190, 1.7273, 1.7…
$ `Maize (tonnes per hectare)`       <dbl> 1.4000, 1.4000, 1.4260, 1.4257, 1.4…
$ `Soybeans (tonnes per hectare)`    <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ `Potatoes (tonnes per hectare)`    <dbl> 8.6667, 7.6667, 8.1333, 8.6000, 8.8…
$ `Beans (tonnes per hectare)`       <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ `Peas (tonnes per hectare)`        <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ `Cassava (tonnes per hectare)`     <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ `Barley (tonnes per hectare)`      <dbl> 1.0800, 1.0800, 1.0800, 1.0857, 1.0…
$ `Cocoa beans (tonnes per hectare)` <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ `Bananas (tonnes per hectare)`     <dbl> NA, NA, NA, NA, NA, NA, NA, NA, NA,…

2.4 Data Cleaning and Preparation

2.4.1 Handling Missing Values

Missing values are common in scientific datasets and need to be addressed before analysis:

Code

# Check for missing values in the penguins dataset
sum(is.na(penguins))

[1] 19

Code

colSums(is.na(penguins))

          species            island    bill_length_mm     bill_depth_mm 
                0                 0                 2                 2 
flipper_length_mm       body_mass_g               sex              year 
                2                 2                11                 0

Code

# Create a complete cases dataset
penguins_complete <- na.omit(penguins)
print(paste("Original dataset rows:", nrow(penguins)))

[1] "Original dataset rows: 344"

Code

print(paste("Complete cases rows:", nrow(penguins_complete)))

[1] "Complete cases rows: 333"

Code

# Replace missing values with the mean for numeric columns
penguins_imputed <- penguins
penguins_imputed$bill_length_mm[is.na(penguins_imputed$bill_length_mm)] <-
  mean(penguins_imputed$bill_length_mm, na.rm = TRUE)
penguins_imputed$bill_depth_mm[is.na(penguins_imputed$bill_depth_mm)] <-
  mean(penguins_imputed$bill_depth_mm, na.rm = TRUE)

# Check if missing values were replaced
sum(is.na(penguins_imputed$bill_length_mm))

[1] 0

Code Explanation

This code demonstrates essential techniques for handling missing values in ecological data:

Missing Value Detection:
- Uses is.na() to identify missing values in the dataset
- sum(is.na()) counts the total number of missing values
- colSums(is.na()) reports missing values per column
- Critical first step in data cleaning
Complete Case Analysis:
- Uses na.omit() to remove rows with any missing values
- Creates a new dataset (penguins_complete) with only complete rows
- Compares the row count before and after removal
- Simple but can lead to significant data loss
Mean Imputation:
- Creates a copy of the original dataset (penguins_imputed)
- Replaces missing values with column means
- Uses logical indexing with is.na() to target only missing values
- Calculates means with na.rm = TRUE to ignore missing values
- Verifies imputation success with another missing value check

Results Interpretation

The output reveals the extent and impact of missing data:

Missing Data Quantity:
- The total number of missing values in the dataset
- The distribution of missing values across columns
- Some columns (like bill measurements) have more missing values than others
Data Loss Impact:
- The original dataset has more rows than the complete cases dataset
- The difference represents the number of incomplete observations
- In ecological studies, this data loss can introduce bias if missingness isn’t random
Imputation Effectiveness:
- After imputation, specific columns no longer contain missing values
- The final check (showing 0) confirms successful imputation
- This approach preserves sample size but may reduce variability

PROFESSIONAL TIP: Handling Missing Values in Ecological Data

When dealing with missing values in ecological datasets:

Understand Missing Data Mechanisms:
- MCAR (Missing Completely At Random): Missingness unrelated to any variables (e.g., equipment failure)
- MAR (Missing At Random): Missingness related to observed variables (e.g., more missing values in certain species)
- MNAR (Missing Not At Random): Missingness related to the missing values themselves (e.g., very small values not detected)
Select Appropriate Handling Methods:
- Complete case analysis: Appropriate for MCAR data with few missing values
- Mean/median imputation: Simple but can underestimate variance
- Multiple imputation: Creates several imputed datasets to account for uncertainty
- Model-based imputation: Uses relationships between variables to predict missing values
- Maximum likelihood: Estimates parameters directly from available data
Document and Report:
- Always report the extent of missing data
- Document your handling approach and rationale
- Consider sensitivity analyses with different approaches
- Acknowledge potential biases introduced by missing data handling

2.4.2 Data Transformation

Often, you’ll need to transform variables to meet statistical assumptions or for better visualization:

Code

# Load the biodiversity dataset
biodiversity <- read_csv("../data/ecology/biodiversity.csv")
glimpse(biodiversity)

Rows: 500
Columns: 24
$ binomial_name     <chr> "Abutilon pitcairnense", "Acaena exigua", "Acalypha …
$ country           <chr> "Pitcairn", "United States", "Congo", "Saint Helena,…
$ continent         <chr> "Oceania", "North America", "Africa", "Africa", "Oce…
$ group             <chr> "Flowering Plant", "Flowering Plant", "Flowering Pla…
$ year_last_seen    <chr> "2000-2020", "1980-1999", "1940-1959", "Before 1900"…
$ threat_AA         <dbl> 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 0, 1…
$ threat_BRU        <dbl> 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0…
$ threat_RCD        <dbl> 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0…
$ threat_ISGD       <dbl> 1, 1, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_EPM        <dbl> 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_CC         <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_HID        <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_P          <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_TS         <dbl> 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_NSM        <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_GE         <dbl> 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ threat_NA         <dbl> 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0…
$ action_LWP        <dbl> 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0…
$ action_SM         <dbl> 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ action_LP         <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ action_RM         <dbl> 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ action_EA         <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0…
$ action_NA         <dbl> 0, 0, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1…
$ red_list_category <chr> "Extinct in the Wild", "Extinct", "Extinct", "Extinc…

Code

# Log transformation of a skewed variable (if available)
if("n" %in% colnames(biodiversity)) {
  biodiversity$log_n <- log(biodiversity$n + 1)  # Add 1 to handle zeros

  # Compare original and transformed
  summary(biodiversity$n)
  summary(biodiversity$log_n)
}

# Standardization (z-score) of penguin measurements
penguins_std <- penguins %>%
  mutate(
    bill_length_std = scale(bill_length_mm),
    flipper_length_std = scale(flipper_length_mm),
    body_mass_std = scale(body_mass_g)
  )

# View the first few rows of the transformed data
head(select(penguins_std, species, bill_length_mm, bill_length_std,
             flipper_length_mm, flipper_length_std), 5)

# A tibble: 5 × 5
  species bill_length_mm bill_length_std[,1] flipper_length_mm
  <chr>            <dbl>               <dbl>             <dbl>
1 Adelie            39.1              -0.883               181
2 Adelie            39.5              -0.810               186
3 Adelie            40.3              -0.663               195
4 Adelie            NA                NA                    NA
5 Adelie            36.7              -1.32                193
# ℹ 1 more variable: flipper_length_std <dbl[,1]>

2.4.3 Creating New Variables

Creating new variables from existing ones is a common data preparation task:

Code

# Create new variables in the penguins dataset
penguins_derived <- penguins %>%
  filter(!is.na(bill_length_mm) & !is.na(bill_depth_mm)) %>%
  mutate(
    bill_ratio = bill_length_mm / bill_depth_mm,
    size_category = case_when(
      body_mass_g < 3500 ~ "Small",
      body_mass_g < 4500 ~ "Medium",
      TRUE ~ "Large"
    )
  )

# View the new variables
head(select(penguins_derived, species, bill_length_mm, bill_depth_mm,
             bill_ratio, body_mass_g, size_category), 5)

# A tibble: 5 × 6
  species bill_length_mm bill_depth_mm bill_ratio body_mass_g size_category
  <chr>            <dbl>         <dbl>      <dbl>       <dbl> <chr>        
1 Adelie            39.1          18.7       2.09        3750 Medium       
2 Adelie            39.5          17.4       2.27        3800 Medium       
3 Adelie            40.3          18         2.24        3250 Small        
4 Adelie            36.7          19.3       1.90        3450 Small        
5 Adelie            39.3          20.6       1.91        3650 Medium

2.5 Data Manipulation with dplyr

The dplyr package provides a powerful grammar for data manipulation:

Code

library(dplyr)

# Filter rows - only Adelie penguins
adelie_penguins <- penguins %>%
  filter(species == "Adelie")
head(adelie_penguins, 3)

# A tibble: 3 × 8
  species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
  <chr>   <chr>              <dbl>         <dbl>             <dbl>       <dbl>
1 Adelie  Torgersen           39.1          18.7               181        3750
2 Adelie  Torgersen           39.5          17.4               186        3800
3 Adelie  Torgersen           40.3          18                 195        3250
# ℹ 2 more variables: sex <chr>, year <dbl>

Code

# Select columns - focus on measurements
penguin_measurements <- penguins %>%
  select(species, island, bill_length_mm, bill_depth_mm, flipper_length_mm, body_mass_g)
head(penguin_measurements, 3)

# A tibble: 3 × 6
  species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
  <chr>   <chr>              <dbl>         <dbl>             <dbl>       <dbl>
1 Adelie  Torgersen           39.1          18.7               181        3750
2 Adelie  Torgersen           39.5          17.4               186        3800
3 Adelie  Torgersen           40.3          18                 195        3250

Code

# Create new variables
penguins_analyzed <- penguins %>%
  mutate(
    bill_ratio = bill_length_mm / bill_depth_mm,
    body_mass_kg = body_mass_g / 1000
  )
head(select(penguins_analyzed, species, bill_ratio, body_mass_kg), 3)

# A tibble: 3 × 3
  species bill_ratio body_mass_kg
  <chr>        <dbl>        <dbl>
1 Adelie        2.09         3.75
2 Adelie        2.27         3.8 
3 Adelie        2.24         3.25

Code

# Summarize data by species
penguin_summary <- penguins %>%
  group_by(species) %>%
  summarize(
    count = n(),
    avg_bill_length = mean(bill_length_mm, na.rm = TRUE),
    avg_bill_depth = mean(bill_depth_mm, na.rm = TRUE),
    avg_body_mass = mean(body_mass_g, na.rm = TRUE)
  ) %>%
  arrange(desc(avg_body_mass))
print(penguin_summary)

# A tibble: 3 × 5
  species   count avg_bill_length avg_bill_depth avg_body_mass
  <chr>     <int>           <dbl>          <dbl>         <dbl>
1 Gentoo      124            47.5           15.0         5076.
2 Chinstrap    68            48.8           18.4         3733.
3 Adelie      152            38.8           18.3         3701.

Code

# Analyze crop yields data
crop_summary <- crops %>%
  filter(!is.na(`Wheat (tonnes per hectare)`)) %>%
  group_by(Entity) %>%
  summarize(
    years_recorded = n(),
    avg_wheat_yield = mean(`Wheat (tonnes per hectare)`, na.rm = TRUE),
    max_wheat_yield = max(`Wheat (tonnes per hectare)`, na.rm = TRUE)
  ) %>%
  arrange(desc(avg_wheat_yield)) %>%
  head(10)  # Top 10 countries by average wheat yield

print(crop_summary)

# A tibble: 10 × 4
   Entity          years_recorded avg_wheat_yield max_wheat_yield
   <chr>                    <int>           <dbl>           <dbl>
 1 Belgium                     19            8.54           10.0 
 2 Netherlands                 58            7.03            9.29
 3 Ireland                     58            6.83           10.7 
 4 United Kingdom              58            6.37            8.98
 5 Denmark                     58            6.18            8.24
 6 Luxembourg                  19            5.98            6.82
 7 Germany                     58            5.89            8.63
 8 Europe, Western             58            5.72            7.88
 9 France                      58            5.65            7.80
10 Northern Europe             58            5.59            7.21

2.6 Exploratory Data Analysis

Before diving into formal statistical tests, it’s essential to explore your data:

Code

# Basic summary statistics
summary(penguins$bill_length_mm)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  32.10   39.23   44.45   43.92   48.50   59.60       2

Code

summary(penguins$flipper_length_mm)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  172.0   190.0   197.0   200.9   213.0   231.0       2

Code

summary(penguins$body_mass_g)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
   2700    3550    4050    4202    4750    6300       2

Code

# Correlation between variables
cor_matrix <- cor(
  penguins %>%
    select(bill_length_mm, bill_depth_mm, flipper_length_mm, body_mass_g),
  use = "complete.obs"
)
print(cor_matrix)

                  bill_length_mm bill_depth_mm flipper_length_mm body_mass_g
bill_length_mm         1.0000000    -0.2350529         0.6561813   0.5951098
bill_depth_mm         -0.2350529     1.0000000        -0.5838512  -0.4719156
flipper_length_mm      0.6561813    -0.5838512         1.0000000   0.8712018
body_mass_g            0.5951098    -0.4719156         0.8712018   1.0000000

Code

# Basic visualization - histogram of bill lengths
library(ggplot2)
ggplot(penguins, aes(x = bill_length_mm)) +
  geom_histogram(binwidth = 1, fill = "skyblue", color = "black") +
  labs(title = "Distribution of Penguin Bill Lengths",
       x = "Bill Length (mm)",
       y = "Frequency") +
  theme_minimal()

Code

# Boxplot of body mass by species
ggplot(penguins, aes(x = species, y = body_mass_g, fill = species)) +
  geom_boxplot() +
  labs(title = "Body Mass by Penguin Species",
       x = "Species",
       y = "Body Mass (g)") +
  theme_minimal() +
  theme(legend.position = "none")

Code

# Scatterplot of bill length vs. flipper length
ggplot(penguins, aes(x = flipper_length_mm, y = bill_length_mm, color = species)) +
  geom_point(alpha = 0.7) +
  labs(title = "Bill Length vs. Flipper Length",
       x = "Flipper Length (mm)",
       y = "Bill Length (mm)") +
  theme_minimal()

Code Explanation

This code demonstrates several key data analysis and visualization techniques:

Summary Statistics:
- summary() provides descriptive statistics for each variable
- Includes min, max, quartiles, mean, and missing values
Correlation Analysis:
- cor() calculates correlation coefficients between variables
- select() chooses specific columns for analysis
- use = "complete.obs" handles missing values
Visualization Components:
- ggplot() creates the base plot
- aes() defines aesthetic mappings
- geom_*() functions add different plot types
- theme_minimal() applies a clean theme

Results Interpretation

The analysis reveals several important insights:

Variable Distributions:
- Bill lengths show a roughly normal distribution
- Body mass varies significantly between species
- Some variables have missing values that need attention
Species Differences:
- The boxplot shows clear species-specific body mass patterns
- Some species show more variation than others
- Potential outliers are visible in the body mass data
Morphological Relationships:
- The scatterplot reveals correlations between bill and flipper lengths
- Species clusters are visible in the morphological space
- Some species show distinct morphological patterns

PROFESSIONAL TIP: Exploratory Data Analysis Best Practices

When conducting exploratory data analysis:

Data Quality:
- Always check for missing values first
- Look for outliers and potential errors
- Verify data types and ranges
Visualization Strategy:
- Start with simple plots (histograms, boxplots)
- Progress to more complex visualizations
- Use appropriate plot types for your data
- Consider colorblind-friendly palettes
Statistical Summary:
- Calculate both descriptive and inferential statistics
- Consider the distribution of your data
- Look for patterns and relationships
- Document any unusual findings

2.7 Summary

In this chapter, we’ve covered the basics of working with data in R:

Understanding different data types and structures
Importing data from various file formats
Cleaning and preparing data for analysis
Creating new variables
Using dplyr for powerful data manipulation
Conducting initial exploratory data analysis

These skills form the foundation for all the analyses we’ll perform in the subsequent chapters. By mastering these basics, you’ll be well-prepared to tackle more complex analytical challenges in various scientific fields.

2.8 Exercises

Load the Palmer Penguins dataset (../data/environmental/climate_data.csv) and create a summary of the number of penguins by species and island.
Calculate the mean and standard deviation of bill length, bill depth, and body mass for each penguin species.
Create a new variable that represents the ratio of flipper length to body mass. Interpret what this ratio might represent biologically.
Create a visualization that shows the relationship between bill length and bill depth, colored by species.
Load the crop yields dataset (../data/agriculture/crop_yields.csv) and analyze trends in wheat yields over time for a country of your choice.
Compare the distributions of body mass between male and female penguins using appropriate visualizations.