Data Exploration, Visualization, and Feature

Engineering using R
Yuhui Zhang, and Raja Iqbal

Basic plotting systems

1. Base graphics: constructed piecemeal. Conceptually simpler and allows plotting to mirror the thought
process.
2. Lattice graphics: entire plots created in a simple function call.
3. ggplot2 graphics: an implementation of the Grammar of Graphics by Leland Wikinson. Combines
concepts from both base and lattice graphics. (Need to install ggplot2 library)
4. Fancier and more telling ones.

A list of interactive visualization in R can be found at: http://ouzor.github.io/blog/2014/11/21/

interactive-visualizations.html

Base plotting system

library(datasets)
## scatter plot
plot(x = airquality$Temp, y = airquality$Ozone)
150
airquality$Ozone

100
50
0

60 70 80 90

airquality$Temp

1
Base plotting system

## par() function is used to specify global graphics parameters that affect all plots in an R session.
## Type ?par to see all parameters
par(mfrow = c(1, 2), mar = c(4, 4, 2, 1), oma = c(0, 0, 2, 0))
with(airquality, {
plot(Wind, Ozone, main="Ozone and Wind")
plot(Temp, Ozone, main="Ozone and Temperature")
mtext("Ozone and Weather in New York City", outer=TRUE)})

Ozone and Weather in New York City

Ozone and Wind Ozone and Temperature
150

150
100

100
Ozone

Ozone
50

50
0

5 10 15 20 60 70 80 90

Wind Temp

Plotting functions (high level)

PHASE ONE: Mount a canvas panel on the easel, and draw the draft. (Initialize a plot.)

• plot(): one of the most frequently used plotting functions in R.

• boxplot(): a boxplot show the distribution of a vector. It is very useful to example the distribution of
different variables.
• barplot(): create a bar plot with vertical or horizontal bars.
• hist(): compute a histogram of the given data values.
• pie(): draw a pie chart.

Remember to use ?plot or str(plot), etc. to check the arguments when you want to make more personalized
plots. A tutorial of base plotting system with more details: http://bcb.dfci.harvard.edu/~aedin/courses/
BiocDec2011/2.Plotting.pdf

Plotting functions (low level)

PHASE TWO: Add more details on your canvas, and make an artwork. (Add more on an existing
plot.)

• lines: adds liens to a plot, given a vector of x values and corresponding vector of y values

2
 • points: adds a point to the plot
• text: add text labels to a plot using specified x,y coordinates
• title: add annotations to x,y axis labels, title, subtitles, outer margin
• mtext: add arbitrary text to margins (inner or outer) of plot
• axis: specify axis ticks

Save your artwork

R can generate graphics (of varying levels of quality) on almost any type of display or printing device. Like:

• postscript(): for printing on PostScript printers, or creating PostScript graphics files.

• pdf(): produces a PDF file, which can also be included into PDF files.
• jpeg(): produces a bitmap JPEG file, best used for image plots.

help(Devices) for a list of them all. Simple example:

## png(filename = 'plot1.png', width = 480, height = 480, units = 'px')

## plot(x, y)
## dev.off()

Example: boxplot and hitogram

## the layout
par(mfrow = c(2, 1), mar = c(2, 0, 2, 0), oma = c(0, 0, 0, 0))
## histogram at the top
hist(airquality$Ozone, breaks=12, main = "Histogram of Ozone")
## box plot below for comparison
boxplot(airquality$Ozone, horizontal=TRUE, main = "Box plot of Ozone")

3
 Histogram of Ozone

0 50 100 150
Box plot of Ozone
airquality$Ozone

0 50 100 150

Lattice plotting system

library(lattice) # need to load the lattice library

set.seed(10) # set the seed so our plots are the same
x <- rnorm(100)
f <- rep(1:4, each = 25) # first 25 elements are 1, second 25 elements are 2, ...
y <- x + f - f * x+ rnorm(100, sd = 0.5)
f <- factor(f, labels = c("Group 1", "Group 2", "Group 3", "Group 4"))
# first 25 elements are in Group 1, second 25 elements are in Group 2, ...
xyplot(y ~ x | f)

−2 −1 0 1 2 −2 −1 0 1 2

Group 1 Group 2 Group 3 Group 4

4
y

−2

−2 −1 0 1 2 −2 −1 0 1 2

Lattice plotting system

Want more on the plot? Customize the panel funciton:

4
xyplot(y ~ x | f, panel = function(x, y, ...) {
# call the default panel function for xyplot
panel.xyplot(x, y, ...)
# adds a horizontal line at the median
panel.abline(h = median(y), lty = 2)
# overlays a simple linear regression line
panel.lmline(x, y, col = 2)
})

Lattice plotting system

−2 −1 0 1 2

Group 3 Group 4
8
6
4
2
0
−2
Group 1 Group 2
y

8
6
4
2
0
−2

−2 −1 0 1 2

Lattice plotting system

Plotting functions * xyplot(): main function for creating scatterplots * bwplot(): box and whiskers plots
(box plots) * histogram(): histograms * stripplot(): box plot with actual points * dotplot(): plot dots on
“violin strings” * splom(): scatterplot matrix (like pairs() in base plotting system) * levelplot()/contourplot():
plotting image data

5
Very useful when we want a lot. . .

pairs(iris) ## iris is a data set in R

2.0 3.0 4.0 0.5 1.5 2.5

6.5
Sepal.Length

4.5
3.5

Sepal.Width
2.0

7
5
Petal.Length

3
1
2.0

Petal.Width
0.5

1.0 2.0 3.0

Species

4.5 6.0 7.5 1 3 5 7 1.0 2.0 3.0

ggplot2

• An implementation of the Grammar of Graphics by Leland Wikinson

• Written by Hadley Wickham (while he was a graduate student as lowa State)
• A “third” graphics system for R (along with base and lattice)
Available from CRAN via install.packages()
web site: http://ggplot2.org (better documentation)
• Grammar of graphics represents the abstraction of graphics ideas/objects
Think “verb”, “noun”, “adjective” for graphics
“Shorten” the distance from mind to page
• Two main functions:
qplot() hides what goes on underneath, which is okay for most operations ggplot() is the core function
and very flexible for doing this qplot() cannot do

qplot function

The qplot() function is the analog to plot() but with many build-in features
Syntax somewhere in between base/lattice
Difficult to be customized (don’t bother, use full ggplot2 power in that case)

6
library(ggplot2) ## need to install and load this library
qplot(displ, hwy, data = mpg, facets = .~drv)

4 f r

40

30
hwy

20

2 3 4 5 6 7 2 3 4 5 6 7 2 3 4 5 6 7
displ

ggplot function

When building plots in ggplot2 (ggplot, rather than using qplot)

The “artist’s palette” model may be the closest analogy
Plots are built up in layers
Step I: Input the data
noun: the data

library(ggplot2) ## need to install and load this library

g <- ggplot(iris, aes(Sepal.Length, Sepal.Width)) ## this would not show you add plot

ggplot function

• Step II: Add layers

adjective: describe the type of plot you will produce.

g + geom_point() + geom_smooth(method = "lm") + facet_grid(. ~ Species)

7
 setosa versicolor virginica
4.5

4.0

3.5
Sepal.Width

3.0

2.5

2.0

5 6 7 8 5 6 7 8 5 6 7 8
Sepal.Length

ggplot function

• Step III: Add metadata and annotation

adjective: control the mapping between data and aesthetics.

g <- g + geom_point() + geom_smooth(method = "lm") + facet_grid(. ~ Species)

g + ggtitle("Sepal length vs. width for different species") + theme_bw() ## verb

Sepal length vs. width for different species

setosa versicolor virginica
4.5

4.0
Sepal.Width

3.5

3.0

2.5

2.0

5 6 7 8 5 6 7 8 5 6 7 8
Sepal.Length

Great documentation

Great documentation of ggplot with all functions in step II and III and demos:
http://docs.ggplot2.org/current/

8
Titanic tragedy data

Reading RAW training data

• Download the data set “Titanic_train.csv” from https://raw.githubusercontent.com/datasciencedojo/

datasets/master/Titanic_train.csv
• Set working directory of R to the directory of the file using setwd()

titanic = read.csv('Titanic_train.csv')

Look at the first few rows

What would be some good features to consider here?

options(width = 110)
head(titanic)

## PassengerId Survived Pclass Name Sex Age SibSp Pa

## 1 1 0 3 Braund, Mr. Owen Harris male 22 1
## 2 2 1 1 Cumings, Mrs. John Bradley (Florence Briggs Thayer) female 38 1
## 3 3 1 3 Heikkinen, Miss. Laina female 26 0
## 4 4 1 1 Futrelle, Mrs. Jacques Heath (Lily May Peel) female 35 1
## 5 5 0 3 Allen, Mr. William Henry male 35 0
## 6 6 0 3 Moran, Mr. James male NA 0
## Ticket Fare Cabin Embarked
## 1 A/5 21171 7.2500 S
## 2 PC 17599 71.2833 C85 C
## 3 STON/O2. 3101282 7.9250 S
## 4 113803 53.1000 C123 S
## 5 373450 8.0500 S
## 6 330877 8.4583 Q

What is the data type of each column?

sapply(titanic,class)

## PassengerId Survived Pclass Name Sex Age SibSp Parch

## "integer" "integer" "integer" "factor" "factor" "numeric" "integer" "integer" "f
## Fare Cabin Embarked
## "numeric" "factor" "factor"

9
Converting class label to a factor

titanic$Survived = factor(titanic$Survived, labels=c("died", "survived"))

titanic$Embarked = factor(titanic$Embarked, labels=c("unkown", "Cherbourg", "Queenstown", "Southampton")
sapply(titanic,class)

## PassengerId Survived Pclass Name Sex Age SibSp Parch

## "integer" "factor" "integer" "factor" "factor" "numeric" "integer" "integer" "f
## Fare Cabin Embarked
## "numeric" "factor" "factor"

str(titanic$Survived)

## Factor w/ 2 levels "died","survived": 1 2 2 2 1 1 1 1 2 2 ...

str(titanic$Sex)

## Factor w/ 2 levels "female","male": 2 1 1 1 2 2 2 2 1 1 ...

Class distribution - PIE Charts

survivedTable = table(titanic$Survived)
survivedTable

##
## died survived
## 549 342

par(mar = c(0, 0, 0, 0), oma = c(0, 0, 0, 0))

pie(survivedTable,labels=c("Died","Survived"))

Died

Survived

10
Is Sex a good predictor?

male = titanic[titanic$Sex=="male",]
female = titanic[titanic$Sex=="female",]
par(mfrow = c(1, 2), mar = c(0, 0, 2, 0), oma = c(0, 1, 0, 1))
pie(table(male$Survived),labels=c("Dead","Survived"), main="Survival Portion Among Men")
pie(table(female$Survived),labels=c("Dead","Survived"), main="Survival Portion Among Women")

Survival Portion Among Men Survival Portion Among Women

Dead
Dead

Survived
Survived

Is Age a good predictor?

Age <- titanic$Age; summary(Age)

## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's

## 0.42 20.12 28.00 29.70 38.00 80.00 177

How about summary segmented by survival

summary(titanic[titanic$Survived=="0",]$Age)

## Min. 1st Qu. Median Mean 3rd Qu. Max.

##

summary(titanic[titanic$Survived=="1",]$Age)

## Min. 1st Qu. Median Mean 3rd Qu. Max.

##

Age distribution by Survival and Sex

11
par(mfrow = c(1, 2), mar = c(4, 4, 2, 2), oma = c(1, 1, 1, 1))
boxplot(titanic$Age~titanic$Sex, main="Age Distribution By Gender",col=c("red","green"))
boxplot(titanic$Age~titanic$Survived, main="Age Distribution By Survival",col=c("red","green"),
xlab="0:Died 1:Survived",ylab="Age")

Age Distribution By Gender Age Distribution By Survival

80

80
60

60
Age
40

40
20

20
0

female male 0 died survived

0:Died 1:Survived

Histogram of Age

hist(Age, col="blue", xlab="Age", ylab="Frequency",

main = "Distribution of Passenger Ages on Titanic")

12
 Distribution of Passenger Ages on Titanic
200
150
Frequency

100
50
0

0 20 40 60 80

Age

Kernel density plot of age

d = density(na.omit(Age)) # density(Age) won't work, need to omit all NAs

plot(d, main = "kernel density of Ages of Titanic Passengers")
polygon(d, col="red", border="blue")

13
 kernel density of Ages of Titanic Passengers
0.030
0.020
Density

0.010
0.000

0 20 40 60 80

N = 714 Bandwidth = 3.226

Comparison of density plots of Age with different Sex

## Package 'sm', version 2.2-5.4: type help(sm) for summary information

14
 Kernel Density Plot of Ages By Sex

female
0.030

male
0.020
Density

0.010
0.000

−20 0 20 40 60 80 100

Age of Passenger

Did Age have an impact on survival?

kernel density of Ages of Titanic Passengers Kernel Density Plot of Ages By Sex Kernel Density Plot of Ages By Survival
0.035

0.035

female died
male survived
0.030

0.030

0.030
0.025

0.025

0.025
0.020

0.020

0.020
Density

Density

Density
0.015

0.015

0.015
0.010

0.010

0.010
0.005

0.005

0.005
0.000

0.000

0.000

0 20 40 60 80 −20 0 20 40 60 80 100 −20 0 20 40 60 80 100

N = 714 Bandwidth = 3.226 Age of Passenger Age of Passenger

Create categorical groupings: Adult vs Child

An example of feature engineering!

## Multi dimensional comparison

Child <- titanic$Age # Isolating age.

15
## Now we need to create categories: NA = Unknown, 1 = Child, 2 = Adult
## Every age below 13 (exclusive) is classified into age group 1
Child[Child<13] <- 1
## Every child 13 or above is classified into age group 2
Child[Child>=13] <- 2

# Use labels instead of 0's and 1's

Child[Child==1] <- "Child"
Child[Child==2] <- "Adult"
# Appends the new column to the titanic dataset
titanic_with_child_column <- cbind(titanic, Child)
# Removes rows where age is NA
titanic_with_child_column <- titanic_with_child_column[!is.na(titanic_with_child_column$Child),]

Fare matters?
Adult Child

500

400

300

female
200

100

0
Fare

500

400

300
male

200

100

0
died survived died survived
Survived
***

16
How about fare, ship class, port embarkation?
unkown Cherbourg Queenstown Southampton

500

400

300
Fare

200

100

1 2 3 1 2 3 1 2 3 1 2 3
Pclass

Diamond data

Overview of the diamond data

data(diamonds) # loading diamonds data set

head(diamonds, 16) # first few rows of diamond data set

## carat cut color clarity depth table price x y z

## 1 0.23 Ideal E SI2 61.5 55 326 3.95 3.98 2.43
## 2 0.21 Premium E SI1 59.8 61 326 3.89 3.84 2.31
## 3 0.23 Good E VS1 56.9 65 327 4.05 4.07 2.31
## 4 0.29 Premium I VS2 62.4 58 334 4.20 4.23 2.63
## 5 0.31 Good J SI2 63.3 58 335 4.34 4.35 2.75
## 6 0.24 Very Good J VVS2 62.8 57 336 3.94 3.96 2.48
## 7 0.24 Very Good I VVS1 62.3 57 336 3.95 3.98 2.47
## 8 0.26 Very Good H SI1 61.9 55 337 4.07 4.11 2.53
## 9 0.22 Fair E VS2 65.1 61 337 3.87 3.78 2.49
## 10 0.23 Very Good H VS1 59.4 61 338 4.00 4.05 2.39
## 11 0.30 Good J SI1 64.0 55 339 4.25 4.28 2.73
## 12 0.23 Ideal J VS1 62.8 56 340 3.93 3.90 2.46
## 13 0.22 Premium F SI1 60.4 61 342 3.88 3.84 2.33
## 14 0.31 Ideal J SI2 62.2 54 344 4.35 4.37 2.71
## 15 0.20 Premium E SI2 60.2 62 345 3.79 3.75 2.27
## 16 0.32 Premium E I1 60.9 58 345 4.38 4.42 2.68

Histogram of carat

17
library(ggplot2)
ggplot(data=diamonds) + geom_histogram(aes(x=carat))

## stat_bin: binwidth defaulted to range/30. Use 'binwidth = x' to adjust this.

7500
count

5000

2500

0 2 4
carat

Density plot of carat

ggplot(data=diamonds) +
geom_density(aes(x=carat),fill="gray50")

18
 1.5

1.0
density

0.5

0.0

0 1 2 3 4 5
carat

Scatter plots (carat vs. price)

ggplot(diamonds, aes(x=carat,y=price)) + geom_point()

19
 15000
price

10000

5000

0 1 2 3 4 5
carat

Carat with colors

g = ggplot(diamonds, aes(x=carat, y=price)) # saving first layer as variable

g + geom_point(aes(color=color)) # rendering first layer and adding another layer

20
 15000

color
D
E
F
price

10000
G
H
I
J

5000

0 1 2 3 4 5
carat

Carat with colors (more details)

g + geom_point(aes(color=color)) + facet_wrap(~color)

21
 D E F

15000

10000

5000

0
G H I
color
D
15000
E
F
price

10000
G
H
5000
I
J
0
J

15000

10000

5000

0
0 1 2 3 4 5
carat

Let’s consider cut and clarity

I1 SI2 SI1 VS2 VS1 VVS2 VVS1 IF

15000

Fair
10000

5000

15000
Good

10000

5000
color
0
D
15000 E
Very Good

F
price

10000
G
5000 H
I
0
J
15000
Premium

10000

5000

15000
Ideal

10000

5000

0
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
carat

22
Your trun!

What is your knowledge of diamond’s price after exploring this data?

Interactive visualization in R - rCharts

• What is rCharts?
Is an R package to create, customize and publish interactive javascript visualizations from R using a
familiar lattice style plotting interface.
• What rCharts can make and how?
Quick start at: http://ramnathv.github.io/rCharts/
• A list of interactive visualization in R can be found at:
http://ouzor.github.io/blog/2014/11/21/interactive-visualizations.html

Tell your story - R Markdown

• R Markdown is an authoring format that enables easy creation of dynamic documents, presentations,
and reports from R.

• It combines the core syntax of markdown (an easy-to-write plain text format) with embedded R code
chunks that are run so their output can be included in the final document.

• Many available output formats including HTML, PDF, and MS Word

• Installation
Use RStudio: already installed
Outside of RStudio: install.packages(“rmarkdown”). A recent version of pandoc (>= 1.12.3) is also
required. See https://github.com/rstudio/rmarkdown/blob/master/PANDOC.MD to install it.

Check out Markdown first

Markdown is a markup language with plain text formatting syntax designed so that it can be
converted to HTML and many other formats using a tool by the same name.

One minute you get the point, and always check the cheat sheets
https://github.com/adam-p/markdown-here/wiki/Markdown-Cheatsheet#lists

23
Then, R Markdown sample code

Download the template:

https://github.com/datasciencedojo/datasets/blob/master/rmarkdownd_template.Rmd

R Markdown

• YAML header
• Edit Markdown, and R chunks
• Run!
RStudio: knitr button
Command line: render(“file.Rmd”)

Cheat sheet of rmarkdown:

http://www.rstudio.com/wp-content/uploads/2015/02/rmarkdown-cheatsheet.pdf

Present your story of Titanic!

Use * Titanic data * Plotting functions in R * R Markdown template * The heart of data explorer
to write your story of Titanic. . .

Hope this is inspiring :)

Titanic

24