Creating and sa v ing more comple x plots DATA VISU AL IZATION IN - - PowerPoint PPT Presentation

▶

Mar 10, 2024 270 likes •515 views

Creating and sa v ing more comple x plots DATA VISU AL IZATION IN R Ron Pearson Instr u ctor Side effects and ret u rn v al u es All R graphics f u nctions are called for their side - e ects The y generate a plot Unlike most f u nctions ,

SLIDE 1

Creating and saving more complex plots

DATA VISU AL IZATION IN R

Ron Pearson

Instructor

SLIDE 2

DATA VISUALIZATION IN R

Side effects and return values

All R graphics functions are called for their side-eects They generate a plot Unlike most functions, they return nothing useful Exception: barplot() function

SLIDE 3

DATA VISUALIZATION IN R

Side effects and return values

library(MASS) tbl <- table(UScereal$shelf) mids <- barplot(tbl, horiz = TRUE, col = "transparent", names.arg = "") mids [,1] [1,] 0.7 [2,] 1.9 [3,] 3.1 text(10, mids, names(tbl), col = "red", font = 2, cex = 2) title("Distribution of cereals by shelf")

SLIDE 4

DATA VISUALIZATION IN R

symbols() shows relations between 3 or more variables

library(MASS) symbols(UScereal$sugars, UScereal$calories, squares = UScereal$shelf, inches = 0.1, bg = rainbow(3)[UScereal$shelf]) title("Cereal calories vs. sugars, coded by shelf")

SLIDE 5

DATA VISUALIZATION IN R

Saving plots as png files

# Divert graphics output to png file png("SavedGraphicsFile.png") # Create the plot symbols(UScereal$sugars, UScereal$calories, squares = UScereal$shelf, inches = 0.1, bg = rainbow(3)[UScereal$shelf]) # Add the title title("Cereal calories vs. sugars, coded by shelf")

SLIDE 6

Let's practice!

DATA VISU AL IZATION IN R

SLIDE 7

Using color effectively

DATA VISU AL IZATION IN R

Ron Pearson

Instructor

SLIDE 8

DATA VISUALIZATION IN R

Limitations of color

Color-blindness: not everyone can see colors Black-and-white reproduction loses all color-coded details Can be overused and lose usefulness

SLIDE 9

DATA VISUALIZATION IN R

Iliinsky & Steele's recommended colors

"Ideally, about six ..." "... hopefully no more than 12 ..." "... and absolutely no more than 20"

SLIDE 10

DATA VISUALIZATION IN R

Iliinsky & Steele's recommended colors

SLIDE 11

DATA VISUALIZATION IN R

Iliinsky & Steele's recommended colors

SLIDE 12

DATA VISUALIZATION IN R

Iliinsky & Steele's recommended colors

SLIDE 13

Let's practice!

DATA VISU AL IZATION IN R

SLIDE 14

Other graphics systems in R

DATA VISU AL IZATION IN R

Ron Pearson

Instructor

SLIDE 15

DATA VISUALIZATION IN R

Why base R?

Flexible Good for exploratory analysis Easy to learn

SLIDE 16

DATA VISUALIZATION IN R

The grid graphics system

Based on the grid package Greater control over low-level graphical details More exible than base graphics Comes at cost of steep learning curve

SLIDE 17

DATA VISUALIZATION IN R

A simple scatterplot in grid

# Get the data and load the grid package library(MASS) x <- UScereal$sugars y <- UScereal$calories library(grid) # This is the grid code required to generate the plot pushViewport(plotViewport()) pushViewport(dataViewport(x, y)) grid.rect() grid.xaxis() grid.yaxis() grid.points(x, y) grid.text("UScereal$calories", x = unit(-3, "lines"), rot = 90) grid.text("UScereal$sugars", y = unit(-3, "lines"), rot = 0) popViewport(2)

SLIDE 18

DATA VISUALIZATION IN R gridBase

Requires familiarity with both graphics systems

SLIDE 19

DATA VISUALIZATION IN R

The lattice graphics system

Built on grid graphics Very good for conditional graphs

SLIDE 20

DATA VISUALIZATION IN R

How does mpg vs horsepower vary by cylinders?

library(MASS) library(lattice) xyplot(MPG.city ~ Horsepower | Cylinders, data = Cars93)

SLIDE 21

DATA VISUALIZATION IN R

The ggplot2 graphics package

Very popular graphics package based on grid graphics The bases for other DataCamp courses Allows us to build complex plots in stages

SLIDE 22

DATA VISUALIZATION IN R

Example with ggplot2

# Sets up plot, but does not display it basePlot <- ggplot(UScereal, aes(x = sugars, y = calories) # Create a simple scatterplot basePlot + geom_point() # Make point shapes depend on shelf variable basePlot + geom_point(shape = as.character(UScereal$shelf)) # Make the points bigger, easier to see basePlot + geom_point(shape = as.character(UScereal$shelf), size = 3)

SLIDE 23

Let's practice!

DATA VISU AL IZATION IN R