Data Visualization in R
Sam Dishong
2024-12-21
Installing and Checking for Packages
packages <- c("ggplot2", "readr", "tidyverse", "dplyr", "ggpubr", "see", "ggExtra", "rmarkdown") # What packages you want
check_install_packages <- function(pkg){
if (!require(pkg, character.only = TRUE)) {
install.packages(pkg, dependencies = TRUE)
library(pkg, character.only = TRUE)}} # Checking and Installing packages
sapply(packages, check_install_packages) # Doing the previous function to all listed packages## $ggplot2
## NULL
##
## $readr
## NULL
##
## $tidyverse
## NULL
##
## $dplyr
## NULL
##
## $ggpubr
## NULL
##
## $see
## NULL
##
## $ggExtra
## NULL
##
## $rmarkdown
## NULLHow to View Dataset and Information
data("USArrests") # Reads in data into your environment
head(USArrests, n = 3) # Views three rows of the data set## Murder Assault UrbanPop Rape
## Alabama 13.2 236 58 21.2
## Alaska 10.0 263 48 44.5
## Arizona 8.1 294 80 31.0?USArrests # Provides background information on the data setThis data set contains statistics for assault, murder, and rape in each of the 50 states. Each variable is the amount of arrests per 100,00 residents. There is also a variable of the percentage of the population living in urban areas. Each value is numberical.
Scatter Plots & Customizations
library(ggplot2) # Loads packages into R session
ggplot(mtcars, aes(x = mpg, y=hp, color = cyl)) + # Makes the plot with the "mtcars" data set and assigns axes and categorical variable
geom_point(size = 2.4, shape = 8) + # Makes the plot a scatter plot with customized points
theme_minimal()+ # Makes the plot a minimal theme
theme(legend.position = "bottom") + # Moves legend to the bottom of the plot
labs(
title = "Effect of Horsepower on Fuel Efficiency",
subtitle = "Catagorized by Number of Cylinders",
x = "Horsepower", y = "Fuel Efficiency (MPG)") # Creates title, subtitle, and axis titles
data("iris") # Reads in data set
ggplot(iris, aes(x = Sepal.Length, y = Petal.Length, color = Species)) + # Makes the plot with the data set and assigns axes and categorical variable
geom_point(size = 3, shape = 17) + # Makes the plot a scatter plot with customized points
theme_classic()+ # Makes the plot a classic theme
theme(legend.position = "right") + # Moves legend to the right of the plot
labs(
title = "Effect of Sepal Length on Petal Length",
subtitle = "Catagorized by species of iris",
x = "Sepal Length (cm)", y = "Petal Length (cm)") # Creates title, subtitle, and axis titles
library(ggplot2) # Loads packages into R session
data("USArrests") # Reads in data set
ggplot(USArrests, aes(x = Murder, y=Assault)) + # Makes the plot with the data set and assigns axes
geom_point(size = 3, shape = 1) + # Makes the plot a scatter plot with customized points
geom_smooth(method = lm, se = TRUE) + # Adds regression line, linear, with standard error
theme_minimal() + # Makes the plot a minimal theme
labs(
title = "Scatter Plot of Assault vs. Murder Rates",
x = "Murder Rate", y = "Assault Rate") # Creates title and axis titles
Line Plots
USArrests$State <- rownames(USArrests) # Adds the state names as a new column in the dataset
USArrests$AverageCrimeRate <- rowMeans(USArrests[, c("Murder", "Assault", "Rape")]) # Calculates the average of the "Murder", "Assault", and "Rape" columns for each state and stores it in a new column AverageCrimeRate
ggplot(data = USArrests, aes(x = State, y = AverageCrimeRate, group = 1)) + # group argument is used to indicate that all the data points should be treated as part of a single group
geom_line(color = "darkgreen") + # Adds a line for the data points
geom_point(color = "red") + # Adds individual data points
theme_minimal() +
theme(axis.text.x = element_text(angle = 90, hjust = 1)) + # Rotates the x-axis labels by 90 degrees and aligns the text right
labs(title = "Line Plot of Average Crime Rate by State",
subtitle = "In Christmas Colors!",
y = "Average Crime Rate")
Violin Plots & Boxplots
library(ggplot2) # Loads packages into R session
CAM <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Violin_Plot_Data.csv")
data_long <- CAM %>% #turns the data into a long format
pivot_longer(
cols = starts_with("Repeat"),
names_to = "Repeat",
values_to = "values")
ggplot(data = data_long, aes(x = F1Performance, y = values, fill = F1Performance)) +
geom_jitter(aes(color = F1Performance), position = position_jitter(0.1), size = 5, alpha = 0.8) +
# colors points to the category, jitters points, sizes them, and makes them transparent
scale_color_manual(values = c("SVMWithDeepShapMaps" = "darkorchid4",
"SVMWithGradCAMMaps" = "chocolate2")) +
# manually colors points
geom_violin(trim = TRUE, alpha = 0.4, size = 2, draw_quantiles = c(0.25, 0.5, 0.75)) +
# adds violin plot with a specific opacity, specific sized outlines, and adds percentage lines
scale_fill_manual(values = c("SVMWithDeepShapMaps" = "darkorchid4",
"SVMWithGradCAMMaps" = "chocolate2")) +
# makes the violin fill colors specific
coord_flip() +
# flips the plot
theme_minimal() +
# adds the minimal theme
theme(axis.title.y = element_blank(), #removes y-axis title
axis.ticks.y = element_blank(), # removes y-axis ticks
axis.text.y = element_blank(), # removes y-axis text
axis.line.x = element_line(color = "black", size = 1.5), # customizes x-axis line
plot.title = element_text(hjust = 0.5, face="bold"), # centers the plot title and makes it bold
panel.grid.major.y = element_blank(), # removes the major grid lines in the y-direction
panel.grid.minor.x = element_blank(), # removes the minor grid lines in the x-direction
panel.grid.major.x = element_line(color = "grey", linetype = "dashed", size = 1.5), # customizes major x grid
legend.position = "none") + # removes legend
stat_summary(fun = median, geom = "point", shape = 21, size = 3, fill = "white", color = "black", stroke = 1.5) +
# calculates medium, adds point, makes it circular, makes it size 3, colors it, and adds the thickness of border
geom_text(aes(x = "SVMWithGradCAMMaps",
label = "SVM + GRAD-CAM++",
y = 0.64),
vjust = -4.5,
color = "darkorange2",
size = 4.5) +
# adds text above violin plot
geom_text(aes(x = "SVMWithDeepShapMaps",
label = "SVM + Deep SHAP",
y = 0.59),
vjust = -4.5,
color = "darkorchid4",
size = 4.5) +
# adds text above violin plot
scale_y_continuous(
limits = c(0.56, 0.74), # adds limits of values on y-axis
breaks = seq(0.56, 0.74, by = 0.02), #adds sequential breaks, by 0.02, between the limits
labels = seq(0.56, 0.74, by = 0.02)) + # adds labels to the y-axis
labs(
title = "Fig. 7. Grad-CAM++ saliency maps capture unique predictive information.", #adds plot title
y = "F1") #adds y-axis title
# NEW PLOT WITH HALF VIOLINS
library(ggplot2) # Loads packages into R session
CAM <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Violin_Plot_Data.csv")
data_long <- CAM %>%
pivot_longer(
cols = starts_with("Repeat"),
names_to = "Repeat",
values_to = "values")
ggplot(data = data_long, aes(x = F1Performance, y = values, fill = F1Performance)) +
geom_jitter(aes(color = F1Performance), position = position_jitter(0.1), size = 5, alpha = 0.8) +
scale_color_manual(values = c("SVMWithDeepShapMaps" = "darkorchid4",
"SVMWithGradCAMMaps" = "chocolate2")) +
geom_violinhalf(trim = TRUE, alpha = 0.4, size = 2, draw_quantiles = c(0.25, 0.5, 0.75)) + # Makes the plot a HALF VIOLIN
scale_fill_manual(values = c("SVMWithDeepShapMaps" = "darkorchid4",
"SVMWithGradCAMMaps" = "chocolate2")) +
coord_flip() +
theme_minimal() +
theme(axis.title.y = element_blank(),
axis.ticks.y = element_blank(),
axis.text.y = element_blank(),
axis.line.x = element_line(color = "black", size = 1.5),
plot.title = element_text(hjust = 0.5, face="bold"),
panel.grid.major.y = element_blank(),
panel.grid.minor.x = element_blank(),
panel.grid.major.x = element_line(color = "grey", linetype = "dashed", size = 1.5),
legend.position = "none") +
stat_summary(fun = median, geom = "point", shape = 21, size = 3, fill = "white", color = "black", stroke = 1.5) +
geom_text(aes(x = "SVMWithGradCAMMaps",
label = "SVM + GRAD-CAM++",
y = 0.64),
vjust = -4.5,
color = "darkorange2",
size = 4.5) +
geom_text(aes(x = "SVMWithDeepShapMaps",
label = "SVM + Deep SHAP",
y = 0.59),
vjust = -4.5,
color = "darkorchid4",
size = 4.5) +
scale_y_continuous(
limits = c(0.56, 0.74),
breaks = seq(0.56, 0.74, by = 0.02),
labels = seq(0.56, 0.74, by = 0.02)) +
labs(
title = "Fig. 7. Grad-CAM++ saliency maps capture unique predictive information.",
y = "F1")
# NEW PLOT WITH VIOLIN + BOX PLOTS
library(ggplot2) # Loads packages into R session
CAM <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Violin_Plot_Data.csv")
data_long <- CAM %>%
pivot_longer(
cols = starts_with("Repeat"),
names_to = "Repeat",
values_to = "values")
ggplot(data = data_long, aes(x = F1Performance, y = values, fill = F1Performance)) +
geom_violin(trim = FALSE, alpha = 0.6, size = 2) +
scale_fill_manual(values = c("SVMWithDeepShapMaps" = "darkorchid4",
"SVMWithGradCAMMaps" = "chocolate2")) +
geom_boxplot(width = 0.15, fill = "white", color = "black", size = 1) + # Adds boxplots on top of the violins for extra information
coord_flip() +
theme_minimal() +
theme(axis.title.y = element_blank(),
axis.ticks.y = element_blank(),
axis.text.y = element_blank(),
axis.line.x = element_line(color = "black", size = 1.5),
plot.title = element_text(hjust = 0.5, face="bold"),
panel.grid.major.y = element_blank(),
panel.grid.minor.x = element_blank(),
panel.grid.major.x = element_line(color = "grey", linetype = "dashed", size = 1.5),
legend.position = "none") +
stat_summary(fun = median, geom = "point", shape = 21, size = 3, fill = "orangered", color = "black", stroke = 1.5) +
geom_text(aes(x = "SVMWithGradCAMMaps",
label = "SVM + GRAD-CAM++",
y = 0.64),
vjust = -4.5,
color = "darkorange2",
size = 4.5) +
geom_text(aes(x = "SVMWithDeepShapMaps",
label = "SVM + Deep SHAP",
y = 0.59),
vjust = -4.5,
color = "darkorchid4",
size = 4.5) +
scale_y_continuous(
limits = c(0.56, 0.74),
breaks = seq(0.56, 0.74, by = 0.02),
labels = seq(0.56, 0.74, by = 0.02)) +
labs(
title = "Fig. 7. Grad-CAM++ saliency maps capture unique predictive information.",
y = "F1")
Density Plots
library(ggplot2) # Loads packages into R session
population_data <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/log_population_data.csv")
ggplot(data = population_data, aes(x = Log10_Current_Population, y = Log10_Past_Population)) +
stat_density_2d(aes(fill = ..level..), geom = "polygon", colour = "white") + # Creates a 2D Density plot
scale_fill_distiller(palette = "YlOrRd", direction = 1) + # Uses custom palette for continuous data and reverses color direction
theme_minimal() +
labs(title = "2D Density Plot of Population Sizes",
x = "Log10(Current population size N0)",
y = "Log10(Past population size N1)")
library(ggplot2) # Loads packages into R session
longevity_data <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/longevity_data.csv")
long <- longevity_data %>% #create a new data frame called "long" that contains all your newly calculated variables
mutate( #mutate tells the program to perform new calculations
log_mass = log10(mass_g), # create a new column called "log_mass" which Log-transforms mass values
log_lifespan = log10(maximum_lifespan_yr)) %>% # create a new column called "log_lifespan" that Log-transforms lifespan value
group_by(order) %>% # this tells it that after "mutate", you are going to start a new function. for each "order" or group of animals
mutate(order_size = n()) #calculate the sample size of each order and put it in a column called "order_size".
#Now you have a sample size for each order, and you have transformed each mass and lifespan value to log form.
long$class <- as.factor(long$class) # Turns class into a categorical variable
long$order_size <- as.numeric(long$order_size) # Turns order_size into a numeric variable
p = ggplot(long, aes(x = log_mass, y = log_lifespan, size = order_size, color = class)) +
geom_point(alpha = 0.3) +
geom_smooth(method = "lm", se = FALSE) +
scale_color_manual(values = c("Aves" = "lightgreen", "Mammalia" = "darkslategrey")) +
labs(title = "Bubble Chart of Longevity and Body Mass",
x = "Log (Body Mass [g])",
y = "Log (Maximum Lifespan [yr])") +
guides(color = "none", size = "none") + # Removes legends
theme_minimal() +
theme(plot.title = element_text(size = 14, face = "bold"),
axis.title.x = element_text(size = 12, face = "bold"), # Customizes axis titles
axis.title.y = element_text(size = 12, face = "bold")) + # Customizes axis titles
annotate("text", x = 5.1, y = 1.9, label = "Aves", size = 5, color = "lightgreen", fontface = "bold") + # Adds annotation
annotate("text", x = 6.5, y = 1.48, label = "Mammals", size = 5, color = "darkslategrey", fontface = "bold") # Adds annotation
ggExtra::ggMarginal(p, type = "density", groupFill = TRUE, alpha = 0.4) # adds the cool density plots to the top and side
Interpretations:
The addition of the density plots makes the density levels across each axis clear, and its also easier to compare aves and mammals.
The body mass/lifespan individual aves is listed as the light green bubbles. and the body mass/lifespan individual mammals is listed as the dark grey bubbles. The averages of individuals for both aves and mammals can be visualized with the light green and dark grey regression lines. Lastly the densities of the individuals of aves and mammals’ body masses can be visualized with the density plots on top of the graph, and the densities of the individuals of aves and mammals’ lifespans can be visualized with the density plots on the right side of the graph.
Both aves and mammals have a positive relationship between longevity and body mass, but the increase is more extreme in aves.
The data is more biased to average lifespans in both aves and mammals, smaller masses in aves, and slightly larger masses in mammals. I believe the smaller masses in aves is necessary to be able to have consistent flight, while it doesn’t matter as much as mammals due to reproduction and ability to get around to feed.
Adding standard error to the regression lines is always an aspect I look for in graphs, and I would have liked it here as well.
# Look at previous plot for annotations
library(ggplot2)
data("airquality")
p = ggplot(airquality, aes(x = Temp, y = Wind, size = 3, color = factor(Month))) +
geom_point(alpha = 0.5) +
scale_color_manual(values = c("5" = "lightblue2",
"6" = "steelblue1",
"7" = "steelblue4",
"8" = "midnightblue",
"9" = "gray18")) +
theme_minimal() +
guides(color = guide_legend(title = "Month"),
size = "none")+
labs(title = "Air Quality Values Over the Span of 5 Months",
x = "Temperature (F)",
y = "Wind (mph)")
ggExtra::ggMarginal(p, type = "density", groupFill = TRUE, alpha = 0.4)
Multiple Panel Plots
library(ggplot2)
data("ChickWeight")
ggplot(data = ChickWeight, aes(x = Time, y = weight, colour = Chick)) +
geom_line(alpha = 0.8) +
geom_smooth(method = "loess", colour = "black", size = 1.2, se = TRUE) + # "Loess" makes curved regression
facet_wrap(~ Diet, ncol = 4) + # Makes panels by variable "Diet" in 4 columns
labs(title = "Chick Growth by Diet Type",
x = "Time (Days)",
y = "Weight (Grams)") +
guides(colour = "none") + # Removes legend
theme_minimal()
library(ggplot2)
library(ggpubr)
data("CO2")
#VIOLIN PLOT
violin <- ggplot(CO2, aes(x = Treatment, y = uptake, fill = Type)) +
geom_violin(trim = FALSE, size = 0.5, alpha = 0.4) +
scale_fill_manual(values = c("Quebec" = "darkslategray3",
"Mississippi" = "orange2")) +
geom_boxplot(width = 0.15, position = position_dodge(width = 0.9)) +
theme_bw() +
labs(y = "CO2 Uptake (μmol/m2sec)",
fill = "Origin") +
theme(axis.title.y = element_text(size = 5))
#LINE PLOT
line <- ggplot(CO2, aes(x = conc, y = uptake)) +
geom_line(aes(color = Treatment)) +
geom_point(aes(color = Treatment)) +
scale_color_manual(values = c("nonchilled" = "darkolivegreen3",
"chilled" = "orchid")) +
facet_wrap(~ Type) + # Multiple panels by "Type" variable
labs(y = "CO2 Uptake (μmol/m2sec)",
x = "Ambient Carbon Dioxide Concentration (mL/L)") +
theme_bw() +
theme(axis.title.y = element_text(size = 5))
#SCATTER PLOT
scatter <- ggplot(CO2, aes(x = conc, y = uptake)) +
geom_point(aes(color = Treatment), shape = 18, size = 4) +
scale_color_manual(values = c("nonchilled" = "darkolivegreen3",
"chilled" = "orchid")) +
geom_smooth(aes(color = Treatment), method = "lm", se = TRUE, size = 1) +
labs(y = "CO2 Uptake (μmol/m2sec)",
x = "Ambient Carbon Dioxide Concentration (mL/L)") +
theme_bw() +
theme(axis.title.y = element_text(size = 5))
#COMBINED PLOT
combined_plots <- ggarrange(line, violin, scatter + rremove("x.text"), # Uses ggarrange to put 3 plots together
labels = c("A", "B", "C"), # Labels panels
ncol = 1, nrow = 3, # Puts panels into one column and 3 rows
heights = c(2, 2.5, 2.2)) # Adjusts heights of panels compared to each other
combined_plot_with_title <- annotate_figure(
combined_plots,
top = text_grob("Cold-Tolerance Effects on CO2 Uptake in Grass Plants", size = 16, face = "bold")) # Adds title to whole plot
combined_plot_with_title # Views final combined plot
Plots From Public Dataset
alcohol_data <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Consumption of alcoholic beverages in Russia 1998-2023.csv")
linep <- ggplot(alcohol_data, aes(x = Year, y = Consumption.of.alcoholic.beverages.L, color = Type)) +
geom_line(size = 1) +
theme_minimal()+
theme(legend.position = "right") +
labs(
title = "Consumption of Alcohol in Russia",
x = "Year", y = "Consumption (L per capita)")
linep
alcohol_data <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Consumption of alcoholic beverages in Russia 1998-2023.csv")
scatterp <- ggplot(alcohol_data, aes(x = Year, y = Consumption.of.alcoholic.beverages.L.Pure, color = Type)) +
geom_point(size = 1, shape = 19) +
geom_smooth(aes(color = Type), method = lm, se = TRUE) +
theme_minimal()+
theme(legend.position = "right") +
labs(
title = "Consumption of Pure Alcohol in Russia",
x = "Year", y = "Consumption of Pure Alcohol (L per capita)")
scatterp
alcohol_data <- read.csv("C:/Users/Sam/OneDrive/Documents/SMCM/Winter R/Consumption of alcoholic beverages in Russia 1998-2023.csv")
alcohol_data$Type <- factor(alcohol_data$Type)
boxp <- ggplot(alcohol_data, aes(x = Type, y = Consumption.of.alcoholic.beverages.L)) +
geom_boxplot(notch = FALSE, fill = "lightblue", color = "black", size = 0.7) +
theme_minimal()+
theme(legend.position = "top") +
labs(
title = "Consumption of Alcohol in Russia",
x = "Types of Alcohol", y = "Consumption of Alcohol (L per capita)")
boxp
combined_plots <- ggarrange(linep, boxp, scatterp + rremove("x.text"), # Uses ggarrange to put 3 plots together
labels = c("A", "B", "C"), # Labels panels
ncol = 1, nrow = 3, # Puts panels into one column and 3 rows
heights = c(2, 2.5, 2.2)) # Adjusts heights of panels compared to each other
combined_plot_with_title <- annotate_figure(
combined_plots,
top = text_grob("Consumption of Alcohol in Russia", size = 16, face = "bold")) # Adds title to whole plot
combined_plot_with_title