knitr::opts_chunk$set(collapse = TRUE, dpi = 300) ## so jittered figs don't always appear to be changed set.seed(1)
Excerpt from the Gapminder data. The main object in this package is the gapminder
data frame or "tibble". There are other goodies, such as the data in tab delimited form, a larger unfiltered dataset, and premade color schemes for the countries and continents.
The gapminder
data frames include six variables, (Gapminder.org documentation page):
| variable | meaning | |:------------|:-------------------------| | country | | | continent | | | year | | | lifeExp | life expectancy at birth | | pop | total population | | gdpPercap | per-capita GDP |
Per-capita GDP (Gross domestic product) is given in units of international dollars, "a hypothetical unit of currency that has the same purchasing power parity that the U.S. dollar had in the United States at a given point in time" -- 2005, in this case.
Package contains two data frames or tibbles:
gapminder
: 12 rows for each country (1952, 1955, ..., 2007). It's a subset of ...gapminder_unfiltered
: more lightly filtered and therefore about twice as many rows.Install gapminder
from CRAN:
install.packages("gapminder")
Or you can install gapminder
from GitHub:
devtools::install_github("jennybc/gapminder")
Load it and test drive with some data aggregation and plotting:
library("gapminder") aggregate(lifeExp ~ continent, gapminder, median) suppressPackageStartupMessages(library("dplyr")) gapminder %>% filter(year == 2007) %>% group_by(continent) %>% summarise(lifeExp = median(lifeExp)) library("ggplot2") ggplot(gapminder, aes(x = continent, y = lifeExp)) + geom_boxplot(outlier.colour = "hotpink") + geom_jitter(position = position_jitter(width = 0.1, height = 0), alpha = 1/4)
country_colors
and continent_colors
are provided as character vectors where elements are hex colors and the names are countries or continents.
head(country_colors, 4) head(continent_colors)
The country scheme is available in this repo as
ggplot2
Provide country_colors
to scale_color_manual()
like so:
... + scale_color_manual(values = country_colors) + ...
library("ggplot2") ggplot(subset(gapminder, continent != "Oceania"), aes(x = year, y = lifeExp, group = country, color = country)) + geom_line(lwd = 1, show_guide = FALSE) + facet_wrap(~ continent) + scale_color_manual(values = country_colors) + theme_bw() + theme(strip.text = element_text(size = rel(1.1)))
# for convenience, integrate the country colors into the data.frame gap_with_colors <- data.frame(gapminder, cc = I(country_colors[match(gapminder$country, names(country_colors))])) # bubble plot, focus just on Africa and Europe in 2007 keepers <- with(gap_with_colors, continent %in% c("Africa", "Europe") & year == 2007) plot(lifeExp ~ gdpPercap, gap_with_colors, subset = keepers, log = "x", pch = 21, cex = sqrt(gap_with_colors$pop[keepers]/pi)/1500, bg = gap_with_colors$cc[keepers])
gapminder
good for?I have used this excerpt in STAT 545 since 2008 and, more recently, in R-flavored Software Carpentry Workshops and a ggplot2
tutorial. gapminder
is very useful for teaching novices data wrangling and visualization in R.
Description:
r nrow(gapminder)
observations; fills a size niche between iris
(150 rows) and the likes of diamonds
(54K rows)r ncol(gapminder)
variablescountry
a factor with r nlevels(gapminder$country)
levelscontinent
, a factor with r nlevels(gapminder$continent)
levelsyear
: going from 1952 to 2007 in increments of 5 yearspop
: populationgdpPercap
: GDP per capitalifeExp
: life expectancyThere are 12 rows for each country in gapminder
, i.e. complete data for 1952, 1955, ..., 2007.
The two factors provide opportunities to demonstrate factor handling, in aggregation and visualization, for factors with very few and very many levels.
The four quantitative variables are generally quite correlated with each other and these trends have interesting relationships to country
and continent
, so you will find that simple plots and aggregations tell a reasonable story and are not completely boring.
Visualization of the temporal trends in life expectancy, by country, is particularly rewarding, since there are several countries with sharp drops due to political upheaval. This then motivates more systematic investigations via data aggregation to proactively identify all countries whose data exhibits certain properties.
Data cleaning code cannot be clean. It's a sort of sin eater.
— Stat Fact (@StatFact) July 25, 2014
The data-raw
directory contains the Excel spreadsheets downloaded from Gapminder in 2008 and 2009 and all the scripts necessary to create everything in this package, in raw and "compiled notebook" form.
If you want to practice importing from file, various tab delimited files are included:
gapminder.tsv
: the same dataset available via library("gapminder"); gapminder
gapminder-unfiltered.tsv
: the larger dataset available via library("gapminder"); gapminder_unfiltered
.continent-colors.tsv
and country-colors.tsv
: color schemesHere in the source, these delimited files can be found:
inst/
sub-directoryOnce you've installed the gapminder
package they can be found locally and used like so:
gap_tsv <- system.file("gapminder.tsv", package = "gapminder") gap_tsv <- read.delim(gap_tsv) str(gap_tsv) gap_tsv %>% # Bhutan did not make the cut because data for only 8 years :( filter(country == "Bhutan") gap_bigger_tsv <- system.file("gapminder-unfiltered.tsv", package = "gapminder") gap_bigger_tsv <- read.delim(gap_bigger_tsv) str(gap_bigger_tsv) gap_bigger_tsv %>% # Bhutan IS here though! :) filter(country == "Bhutan")
Gapminder's data is released under the Creative Commons Attribution 3.0 Unported license. See their terms of use.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.