• Learn how to:
  • Explore data visualy
  • Create layered graphs w/ ggplot2
• Readings
  • R4DS: ch. 3, 7
  • Data Visualization cheatsheet

• Communicate information graphically through plots, charts, maps, etc.

• High-level plot description based on layered grammar of graphics
  • Create plots by overlaying geometric objects on same coordinates

• R package ggplot2 implements such grammar for creating plots in R

• Basic ingredients
  • Data
  • Aesthetic Mappings
  • Geometric Objects (lines, points, text, etc)
• Extras
  • Statistics
  • Scales
  • Facets
  • Other adjustments (coordinates, annotations, positions)

• Geometric objects convey information through their aesthetics (i.e. perceived characteristics)

• Data variable(s) can be mapped to one or more aesthetics

• Common aesthetic mappings
  • Location: x/y coordinates
  • Appearance: size, color, shape, etc

• Gapminder Foundation is non-profit organization promoting world-wide social, economic and environmental development through the use of data and statistics
  • Excellent source of information about state of the world
• gapminder package provides excerpt of available data

library(gapminder); glimpse(gapminder)
## Observations: 1,704
## Variables: 6
## $ country   <fct> Afghanistan, Afghanistan, Afghanistan, Afghanistan, ...
## $ continent <fct> Asia, Asia, Asia, Asia, Asia, Asia, Asia, Asia, Asia...
## $ year      <int> 1952, 1957, 1962, 1967, 1972, 1977, 1982, 1987, 1992...
## $ lifeExp   <dbl> 28.801, 30.332, 31.997, 34.020, 36.088, 38.438, 39.8...
## $ pop       <int> 8425333, 9240934, 10267083, 11537966, 13079460, 1488...
## $ gdpPercap <dbl> 779.4453, 820.8530, 853.1007, 836.1971, 739.9811, 78...

• Basic syntaxt
  ggplot( data, aes(...) ) + geom_*() + ...

gap07 = gapminder %>% filter( year == "2007")
ggplot(gap07, aes(x = continent)) + geom_bar() 

• Barplot (geom_bar)
• Histogram (geom_histogram)
• Boxplot (geom_boxplot)
• Scatterplot (geom_point)
• Line plot (geom_line)

• Compare values of discrete variable
  • Often used for counts/sums across groups

gap07 %>% group_by(continent) %>% 
  summarise( pop = sum( as.numeric(pop) ) ) %>% 
  ggplot(aes(x = continent, y = pop)) + geom_bar(stat = "identity")

• Describe distribution of continuous variable

ggplot(gap07, aes(x = lifeExp)) + geom_histogram(bins = 20)

• Compare distributions of continuous variable across levels of discrete variable

ggplot(gap07, aes(x = continent, y = lifeExp)) + geom_boxplot()

• Describe continuous distribution using 5-point-summary

• Look at relationship between two continuous variables

ggplot(gap07, aes(x = gdpPercap, y = lifeExp)) + geom_point()

• For data with serial structure (e.g. time series)

gapminder %>% group_by(year) %>% 
  summarise( world_pop = sum(as.numeric(pop))) %>% 
  ggplot(aes(x = year, y = world_pop)) + geom_line()

• Original data can be transformed through stat function

ggplot(gap07, aes(x = continent, y = pop) ) + 
  stat_summary(fun.y = "sum", geom = "bar") 

• Scales control range of aesthetic mappings.

ggplot(gap07, aes(x = continent, y = gdpPercap) ) + 
  geom_boxplot() + scale_y_log10()

• Create grid of sub-plots, one for each level of a variable

ggplot(gap07, aes(x = gdpPercap, y = lifeExp)) + geom_point() +
  facet_wrap( facets = ~ continent)

• Using positional adjustments for bars

gapminder %>% filter( year > 1985 ) %>% 
  mutate( year = as.factor(year), pop = as.numeric(pop) ) %>% 
  group_by(year, continent) %>% summarise(pop = sum(pop) ) %>% 
  ggplot(aes(x = continent, y = pop, fill = year)) + 
  geom_bar(stat = "identity", position = "dodge")

• Manipulate (fix, flip, transform) coordinates

ggplot(gap07, aes(x = continent, y = pop) ) + 
  stat_summary(fun.y = "sum", geom = "bar") + 
  coord_flip()