Inspiring people:
Hadley Wickham: grammar of graphics
Hans Rosling: Gapminder
Gapminder World - Wealth & Health of Nations
ggplot
library(dplyr)
library(ggplot2)
library(gapminder)
# preview data
gapminder
# get range of available data
summary(gapminder)
# setup dataframe
g = gapminder %>%
filter(year==2007) %>% # most recent year
mutate(pop_m = pop/1e6) # population, millions
# plot scatterplot of most recent year
s = ggplot(g, aes(x=gdpPercap, y=lifeExp)) +
geom_point()
s
# add aesthetic of size by population
s = s +
aes(size=pop_m)
s
# add aesthetic of color by continent
s = s +
aes(color=continent)
s
# add title, update axes labels
s = s +
ggtitle('Health & Wealth of Nations for 2007') +
xlab('GDP per capita ($/year)') +
ylab('Life expectancy (years)')
s
# label legend
s = s +
scale_colour_discrete(name='Continent') +
scale_size_continuous(name='Population (M)')
s
Your Turn
Now with country emissions datasets…
# boxplot by continent
b = ggplot(g, aes(x=continent, y=lifeExp)) +
geom_boxplot()
b
# match color to continents, like scatterplot
b = b +
aes(fill=continent)
b
# drop legend, add title, update axes labels
b = b +
theme(legend.position='none') +
ggtitle('Life Expectancy by Continent for 2007') +
xlab('Continent') +
ylab('Life expectancy (years)')
b
Your Turn: Make a similar plot but for gdpPercap
. Be sure to update the plot’s aesthetic, axis label and title accordingly.
plotly
library(plotly) # install.packages('plotly')
# scatterplot (Note: key=country shows up on rollover)
s = ggplot(g, aes(x=gdpPercap, y=lifeExp, key=country)) +
geom_point()
ggplotly(s)
# boxplot
ggplotly(b)
Your Turn: Expand the interactive scatterplot to include all the other bells and whistles of the previous plot in one continuous set of code (no in between setting of s).
library(explodingboxplotR) # devtools::install_github('timelyportfolio/explodingboxplotR')
exploding_boxplot(g,
y = 'lifeExp',
group = 'continent',
color = 'continent',
label = 'country')
The googleVis
package ports most of the Google charts functionality.
For every R chunk must set option results='asis'
, and once before any googleVis plots, set op <- options(gvis.plot.tag='chart')
.
suppressPackageStartupMessages({
library(googleVis) # install.packages('googleVis')
})
op <- options(gvis.plot.tag='chart')
m = gvisMotionChart(
gapminder %>%
mutate(
pop_m = pop / 1e6,
log_gdpPercap = log(gdpPercap)),
idvar='country',
timevar='year',
xvar='log_gdpPercap',
yvar='lifeExp',
colorvar='continent',
sizevar='pop_m')
plot(m)
Your Turn: Repeat the motion chart with the country having the highest gdpPercap
filtered out.
Thematic maps tmap
:
library(tmap) # install.packages('tmap')
# load world spatial polygons
data(World)
# inspect values in World
World@data %>% tbl_df()
# gapminder countries not in World. skipping for now
g %>%
anti_join(World@data, by=c('country'='name')) %>%
arrange(desc(pop))
# World countries not in gapminder. skipping for now
World@data %>%
anti_join(g, by=c('name'='country')) %>%
arrange(desc(pop_est)) %>%
select(iso_a3, name, pop_est)
# join gapminder data to World
World@data = World@data %>%
left_join(g, by=c('name'='country'))
# make map
m = tm_shape(World) +
tm_polygons('lifeExp', palette='RdYlGn', id='name', title='Life expectancy (years)', auto.palette.mapping=F) +
tm_style_gray() + tm_format_World()
m
# show interactive map
tmap_leaflet(m)