HomeWork3

This assignment is done with the help of RMarkdown

tinytex::install_tinytex() library(tinytext)

Q1: 1. Modify the following plot so that you get one boxplot per integer value of displ.

ggplot(mpg, aes(displ, cty)) + geom_boxplot()

library(ggplot2)
ggplot(mpg, aes(displ, cty)) +
geom_boxplot()

## Warning: Continuous x aesthetic -- did you forget aes(group=...)?

I think here we missed the group displ part as we can see in the warning.

library(ggplot2)
ggplot(mpg, aes(displ, cty)) +
  geom_boxplot(aes(group = displ))

Q2: When illustrating the difference between mapping continuous and discrete colors to a line, the discrete example needed aes(group = 1). Why? What happens if that is omitted? What’s the difference between aes(group = 1) and aes(group = 2)? Why?

In ggplot we need to manuyally specify the grouping. Specifying group = 1 implies that we are using a single line connecting all the points.

e.g.for using the group=1 command

library(ggplot2)
ggplot(data = diamonds) + geom_bar(mapping = aes(x = cut, y = ..prop.., group = 1))

without using the group=1 command. It gives the wrong representation of the plot and matches all values to 1.

library(ggplot2)
library(ggplot2)
ggplot(data = diamonds) + geom_bar(mapping = aes(x = cut, y = ..prop..))

group =1 helps creates a dummy variable to counter effect the denominator issue while calculating proportions. There will not be any difference for group =1 and group =2 unless 2 is a category.

library(ggplot2)
ggplot(data = diamonds) + geom_bar(mapping = aes(x = cut, y = ..prop.., group = 2))

## Q3: How many bars are in each of the following plots? (Hint: try adding an outline around each bar with color = "white")

library(ggplot2)
ggplot(mpg, aes(drv)) +
geom_bar()

library(ggplot2)
ggplot(mpg, aes(drv, fill = hwy, group = hwy)) +
geom_bar()

library(ggplot2)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

mpg2 <- mpg %>% arrange(hwy) %>% mutate(id = seq_along(hwy))
ggplot(mpg2, aes(drv, fill = hwy, group = id)) +
geom_bar()

All plots has 3 bars

Q4: What happens when you put the geom smooth() function before geom point() instead of after it? What does this tell you about how the plot is drawn?

geom_smooth: adds standard error bands to plot lines. The function geom_point() makes scatter plots.

Plotting population vs year with data from gapminder

using geom_point() before geom_smooth

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_point()+
  geom_smooth()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

using geom_point() after geom_smooth

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_smooth()+
  geom_point()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

There is no difference on data data smoothing line neither location of points. The only difference I can see is overlay properties. In the first case where geom_point() is used before geom_smooth(), geom_point() are behind the geom_smooth() layer. This tells us that ggplot creates layers in sequence on top of each other.

Q5) Try some alternative scale mappings. Besides scale x log10() you can try scale x sqrt() and scale x reverse(). There are corresponding functions for y-axis transformations. Just write y instead of x. What sort of effect they have on the plot? When should these scales be used?

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_point()+
  geom_smooth()+
scale_x_log10()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_point()+
  geom_smooth()+
scale_x_sqrt()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_point()+
  geom_smooth()+
scale_x_reverse()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

library(ggplot2)
library("gapminder")
ggplot(gapminder, aes(pop,year)) + 
  geom_point()+
  geom_smooth()+
scale_y_reverse()

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

We should use log scales when we have skewness towards large values and when we want to show percent change or multiplicative factors. Sometime reversing axis makes more sense for certain types of plots and data.

##Q6)Experiment with different ways to facet the data. What happens when you try plotting population and per capita GDP while faceting on year, or even on country? Experimentwith the height and width of the figure.

population and per capita GDP while faceting on year:

library(ggplot2)
library(gapminder)
library(dplyr)
ggplot(gapminder) +
geom_point(aes(x = gdpPercap, y = pop, color = continent)) +
scale_x_log10() +
facet_wrap(~ year)

population and per capita GDP while faceting on country:

library(ggplot2)
library(gapminder)
library(dplyr)
ggplot(gapminder) +
geom_point(aes(x = gdpPercap, y = pop, color = continent)) +
scale_x_log10() +
facet_wrap(~ country)

population and per capita GDP while faceting on country:

library(ggplot2)
library(gapminder)
library(dplyr)
ggplot(gapminder) +
geom_point(aes(x = gdpPercap, y = pop, color = continent)) +
scale_x_log10() +
facet_wrap(~ year)+
theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5))

Faceting on year is not feasible as there are too many countries for one plot. Faceting with year is managable.

Experiments with height and width of figure:

I am not sure for fixing width and height as we have to submit assignment as an .RMD file. But we can export the graphs in desired width and height in R. e.g. png(filename=“DataVizClass.png”, width=600, height=600)

Or we can play with axis to change height and width of plot:

library(ggplot2)
library(gapminder)
library(dplyr)
ggplot(gapminder) +
geom_point(aes(x = gdpPercap, y = pop, color = continent)) +
scale_y_log10() +
facet_wrap(~ year)

library(ggplot2)
library(gapminder)
library(dplyr)
ggplot(gapminder) +
geom_point(aes(x = gdpPercap, y = pop, color = continent)) +
facet_wrap(~ year)

Q7) What is the difference between a formula written as facet grid(sex ∼race) versus one written as facet grid(∼sex + race).

facet_grid(sex ~ race) will facet the relationship by sex and race of the respondent.

install.packages(“socviz”)

library(ggplot2)
library(socviz)
p <- ggplot(data = gss_sm,
            mapping = aes(x = age, y = childs))
p + geom_point(alpha = 0.2) +
    geom_smooth() +
    facet_grid(sex ~ race)

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

## Warning: Removed 18 rows containing non-finite values (stat_smooth).

## Warning: Removed 18 rows containing missing values (geom_point).

library(ggplot2)
library(socviz)
p <- ggplot(data = gss_sm,
            mapping = aes(x = age, y = childs))
p + geom_point(alpha = 0.2) +
    geom_smooth() +
    facet_grid(sex ~ race)

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

## Warning: Removed 18 rows containing non-finite values (stat_smooth).

## Warning: Removed 18 rows containing missing values (geom_point).

library(ggplot2)
library(socviz)
p <- ggplot(data = gss_sm,
            mapping = aes(x = age, y = childs))
p + geom_point(alpha = 0.2) +
    geom_smooth() +
    facet_grid(~sex + race)

## `geom_smooth()` using method = 'gam' and formula 'y ~ s(x, bs = "cs")'

## Warning: Removed 18 rows containing non-finite values (stat_smooth).

## Warning: Removed 18 rows containing missing values (geom_point).

library(ggplot2)