Lab04

R Programming (EDA 2)

Adapted from R4DS Chapter 7: Exploratory Data Analysis

Prerequisite: use preload function to get tidyverse and descr load them

library(tidyverse)

## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──

## ✓ ggplot2 3.3.5     ✓ purrr   0.3.4
## ✓ tibble  3.1.6     ✓ dplyr   1.0.8
## ✓ tidyr   1.2.0     ✓ stringr 1.4.0
## ✓ readr   2.1.2     ✓ forcats 0.5.1

## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()

library(descr) # Describe attributes of objects/variables

Explore the diamonds dataset

?diamonds

Examine the variables using the class function

class(cut)

## [1] "function"

sapply(diamonds,class)

## $carat
## [1] "numeric"
## 
## $cut
## [1] "ordered" "factor" 
## 
## $color
## [1] "ordered" "factor" 
## 
## $clarity
## [1] "ordered" "factor" 
## 
## $depth
## [1] "numeric"
## 
## $table
## [1] "numeric"
## 
## $price
## [1] "integer"
## 
## $x
## [1] "numeric"
## 
## $y
## [1] "numeric"
## 
## $z
## [1] "numeric"

Frequency table from descr package

attach(diamonds)
freq(cut,plot=F) # can add a plot by default

## cut 
##           Frequency Percent Cum Percent
## Fair           1610   2.985       2.985
## Good           4906   9.095      12.080
## Very Good     12082  22.399      34.479
## Premium       13791  25.567      60.046
## Ideal         21551  39.954     100.000
## Total         53940 100.000

Frequency table, alternative method

with(diamonds, {freq(cut, plot=T)})

## cut 
##           Frequency Percent Cum Percent
## Fair           1610   2.985       2.985
## Good           4906   9.095      12.080
## Very Good     12082  22.399      34.479
## Premium       13791  25.567      60.046
## Ideal         21551  39.954     100.000
## Total         53940 100.000

Create tibble object out of diamonds dataset

diam_tb=as_tibble(diamonds)
class(diam_tb) # How is this different from a data frame?

## [1] "tbl_df"     "tbl"        "data.frame"

Plot the bar chart using ggplot2

ggplot(data = diamonds) +
  geom_bar(mapping = aes(x = cut)) 

Better plot with percentage on y axis

ggplot(data = diamonds,aes(cut)) +
  geom_bar(mapping = aes(y = (..count..)/sum(..count..))) + theme_bw() +
  scale_y_continuous(labels=scales::percent) +
  ylab("Percent")

Exercise 1

Save the chart into PNG, PDF and SVG formats

##pdf

setwd("/Users/dasha/Desktop")
pdf("barplot.pdf")
library("ggplot2")
ggplot(data = diamonds,aes(cut)) +
  geom_bar(mapping = aes(y = (..count..)/sum(..count..))) + theme_bw() +
  scale_y_continuous(labels=scales::percent) +
  ylab("Percent")
dev.off()

## quartz_off_screen 
##                 2

##png

png("barplot.png")
ggplot(data = diamonds,aes(cut)) +
  geom_bar(mapping = aes(y = (..count..)/sum(..count..))) + theme_bw() +
  scale_y_continuous(labels=scales::percent) +
  ylab("Percent")
dev.off()

## quartz_off_screen 
##                 2

##svg

svg("barplot.svg")
ggplot(data = diamonds,aes(cut)) +
  geom_bar(mapping = aes(y = (..count..)/sum(..count..))) + theme_bw() +
  scale_y_continuous(labels=scales::percent) +
  ylab("Percent")
dev.off()

## quartz_off_screen 
##                 2

What are the differences among all these formats?

A PNG file is pixel-based image format. If scaled up, it becomes grainy/pixelated and if too small, it can become indistinct.

A SVG file is vector-based. This means that they use mathematical algorithms to display images, which you can then scale to any size without negatively impacting their quality.

A PDF is also vector-based but of larger file size than SVGs.

Plot another variable carat

ggplot(data = diam_tb) +
  geom_histogram(mapping = aes(x = carat), binwidth = 0.5)

Exercise 2

Can you improve the visualization of this chart?

ggplot(data = diam_tb) +
  geom_histogram (mapping = aes(x = carat), fill="#FF6666", binwidth = 0.03) +
  labs(title="Quantity of diamonds according to weight",x="Diamond Weight", y = "Count")+
  theme_classic()

What is the difference between barchart and histogram?

Barchart is for representing categorical variables and historgram for continuous variables

Subsetting the dataset

smaller <- diamonds %>% 
  filter(carat < 3)

Histogram of smaller dataset

ggplot(data = smaller, mapping = aes(x = carat)) +
  geom_histogram(binwidth = 0.1)

Polygon

ggplot(data = smaller, mapping = aes(x = carat, colour = cut)) +
  geom_freqpoly(binwidth = 0.1) + theme_bw() 

Exercise 3

Can you change colors?

Hint: use the RcolorBrewer package

library(RColorBrewer)

Continuous variables

ggplot(data = faithful) + 
  geom_point(mapping = aes(x = eruptions, y = waiting), color='blue')+
  labs(title="Eruptions by time of waiting",x="Eruptions", y = "Waiting")+
  theme_classic()

Exercise 4

What method you will use to analyze:

Dependent variable: continuous, Independent variable: discrete

ANOVA

Dependent variable: discrete, Independent variable: continuous

Logistic regression or logit for binary response and Poisson regression for a count-based DV

Dependent variable: continuous, Independent variable: continuous

Linear regression