Simple example of creating a missingness map with country-year data

missing_map.R

# ---------------------------------------------------------------------------- #
# Create a missingness map to examine patterns of missing values across variables
# Christopher Gandrud
# MIT LICENSE
# ---------------------------------------------------------------------------- #

# Load required packages
library(WDI)
library(dplyr)
library(Amelia)

# Download 5 variables from the World Bank Development Indicators ----------
indicators <- c('EN.ATM.METH.KT.CE', 'EG.USE.ELEC.KH.PC', 'EN.ATM.CO2E.PC', 
                'SP.POP.GROW', 'EG.USE.COMM.CL.ZS')

wdi <- WDI(indicator = indicators, start = 1990, end = 2014)

# Rename indicators and drop iso2c variable
wdi <- wdi %>% rename(methane_emissions = EN.ATM.METH.KT.CE) %>% 
                rename(electricity_use = EG.USE.ELEC.KH.PC) %>%
                rename(co2_emissions = EN.ATM.CO2E.PC) %>% 
                rename(population_growth = SP.POP.GROW) %>% 
                rename(alternative_energy = EG.USE.COMM.CL.ZS) %>%
                select(-iso2c)

# Subset the data (just to make the illustration easier to understand)
wdi_sub <- wdi %>% filter(country %in% c('Afghanistan', 'Belgium', 'China', 
                                       'Colombia', 'France', 'Germany', 
                                       'South Africa', 'Zambia'))

# Create missingness map ---------------------------------------------------
missmap(obj = wdi_sub,
        csvar = 'country',  # country identifying variable 
        tsvar = 'year',     # time identifying variable
        y.cex = 0.75,       # decrease y-axis labels so they fit
        x.cex = 0.55        # decrease x-axis labels so they fit
)