Data Analysis by R

ggplot2

ggplot2 is the software made for Graphical Analysis in R. Today there is for Python. The code in this page is for R.

Common Code

This example is needed the code starting "ggplot".

library(ggplot2)
setwd("C:/Rtest")
Data <- read.csv("Data.csv", header=T) # Read data

Short Pass for the Graph

Graph for the analysis the relationship 2 variables
>>>Line Graph
>>>Stratified Line Graph
>>>Same Space Line Graph
>>>Basic Scatter Plot
>>>Stratified Scatter Plot
>>>Regression Line
>>>Scatter Plot for Words

Histgram
>>>Basic Histgram
>>>Stratified Histgram
>>>Good Range Histgram

Graph for 1 Variable
>>>Graph for 1 Variable Stratified by 1 Variable
>>>Graph for 1 Variable Stratified by 2 Variables
>>>Graph for 1 Variable Stratified by 3 Variables

Bar Plot
>>>Basic Bar Plot
>>>Stratified Bar Plot
>>>Frequency Plot

Adjusting the Graph
>>>Angle of Label

Graph for the analysis the relationship 2 variables

There is Line Graph as Graph for Change .

Line Graph

After Common Code

ggplot(Data, aes(x=X1,y=Y1, group=1)) + geom_line() + geom_point()# Line Graph

Stratified Line Graph

This code is imaged that stratified data is not in same variable.

library(ggplot2)
setwd("C:/Rtest")
Data <- read.csv("Data.csv", header=T)
library(tidyr)
Data_long <- tidyr::gather(Data, key="Yno", value = Ys, -X1) # Pile up the data
ggplot(Data_long, aes(x=X1,y=Ys, colour=Yno)) + geom_line() + geom_point()# Stratified Line Graph

facet is used to separate the are of graphs.
ggplot(Data_long, aes(x=X1,y=Ys)) + geom_line() + geom_point() + facet_wrap(~Yno,nrow=1000)

Y range is adjusted for each graph.
ggplot(Data_long, aes(x=X1,y=Ys)) + geom_line() + geom_point() + facet_wrap(~Yno,scales="free",nrow=1000)

The effect of adjusting the Y range is clear if units or orders of each variable are different. Graphs below are made to check this effect.

ggplot(Data_long, aes(x=X1,y=Ys)) + geom_line() + geom_point() + facet_wrap(~Yno,scales="free") # Make many line graghs that of Y ranges are different

Same Space Line Graph

In the graphs above, space of plots of X axis are not same because X axis data is number type data. If we want same space, changing data type into factor data is one of the solution.
Data$X1 <-factor(Data$X1) # Changing data type into factor data
ggplot(Data, aes(x=X1,y=Y1, group=1)) + geom_line() + geom_point()

The other solution is changing X1 into numbers of columns.
Data$X1 <-as.numeric(row.names(Data)) # changing X1 into numbers of columns
ggplot(Data, aes(x=X1,y=Y1)) + geom_line() + geom_point()

Stratified Scatter Plot

ggplot(Data, aes(x=X1, y=Y1)) + geom_point(aes(colour=C1, shape=C1))

Adjust Range

ggplot(Data, aes(x=X1, y=Y1)) + geom_point(aes(colour=C1, shape=C1)) + coord_cartesian(xlim=c(0,20),ylim=c(0,20))

Regression Line

ggplot(Data, aes(x=X1, y=Y1)) + geom_point(aes(colour=C1, shape=C1)) + geom_smooth(method = "lm",aes(fill = C1))

If without "aes(fill = C1)", regression line is made for all data not for each category.

Scatter Plot for Words

Scatter Plot for Words is also made.

ggplot(Data, aes(x=X1, y=Y1,label=C1)) + geom_text() # Scatter Plot for Words

Histgram

Histgram is made

Histgram is one of the Graph for 1 Variable . But I separate the introduction because making process is different.

Basic Histgram

ggplot(Data, aes(x=Y1)) + geom_histogram()

Stratified Histgram

ggplot(Data, aes(x=Y1)) + geom_histogram() + facet_grid(C1~.)

ggplot(Data, aes(x=Y1)) + geom_histogram() + facet_grid(C1+C2~.)

Good Range Histgram

Default setting od ggplot2 make 30 section for data range.

Example below is made to change sections for easy understanding of data.

Data[,'Y1r']<-trunc(Data[,'Y1']/10)*10# unit is 10
ggplot(Data, aes(x=Y1r)) + geom_histogram(binwidth = 10)

Example to change the section Data[,'Y1r']<-trunc(Data[,'Y1']) # unit is 1
Data[,'Y1r']<-trunc(Data[,'Y1']/0.5)*0.5 # unit is 0.5

Graph for 1 Variable

Graph for 1 Variable Stratified by 1 Variable

Making 1-Dimension Scatter Plot .

ggplot(Data, aes(x=C1, y=Y1)) + geom_point()

ggplot(Data, aes(x=C1, y=Y1)) + geom_jitter(size=1, position=position_jitter(0.1))# 1-Dimension Jitter Plot

Making Box Plot .

ggplot(Data, aes(x=C1, y=Y1)) + geom_boxplot()
Graph for 1 Variable Stratified by 2 Variables

ggplot(Data, aes(x=C1, y=Y1)) + geom_jitter(size=1, position=position_jitter(0.1)) +facet_grid(.~C2)

Example of many categories.

"facet_grid" puts graphs on a line. "facet_wrap" puts graphs in a good balance.

ggplot(Data, aes(x=C2, y=Y1)) + geom_point() +facet_wrap(~C1)

Graph for 1 Variable Stratified by 3 Variables

ggplot(Data, aes(x=C1, y=Y1)) + geom_jitter(size=1, position=position_jitter(0.1)) +facet_grid(.~C2+C3)

ggplot(Data, aes(x=C1, y=Y1)) + geom_jitter(size=1, position=position_jitter(0.1)) +facet_grid(C2~C3)

Bar Plot

Basic Bar Plot

Making Bar Plot

ggplot(Data, aes(x=C1, y=Y1)) + geom_bar(stat = "identity")

Stratified Bar Plot

ggplot(Data, aes(x=C1, y=Y1)) + geom_bar(stat="identity", aes(fill=C1)) + facet_grid(. ~ C2)

Frequency Plot

"C1" and "C2" are categorical data.

ggplot(Data, aes(x=C1)) + geom_bar() + facet_grid(. ~ C2)

Adjusting the Graph

Angle of Label

Changing the angle of label/

ggplot(Data, aes(x=C1, y=Y1)) + geom_bar(stat = "identity") + theme(axis.text = element_text(angle = 90))

ggplot(Data, aes(x=C1, y=Y1)) + geom_boxplot() + theme(axis.text = element_text(angle = 90))