Click the Original, Code and Reconstruction tabs to read about the issues and how they were fixed.

Original


Source: Carbon Dioxide Information Analysis Center; Research Institute For Environment, Energy and Economics; ND-Gain Index


Objective

To inform a general audience interested in climate change issues, the contrast of nations that are the higher CO2 emittters with those nations that are most vulnerable to climate change. Both groups are almost mutually exclusive.

The visualisation chosen had the following three main issues:

  • Perceptual issue #1 - accuracy of size of either emissions or vulnerability is lost through the use of area (disc) and the positioning there of (including labelling of nations) on an arc.
  • Perceptual issue #2 - conveying visually of the two features, emissions and vulnerability, independently (nation is one or the other), when they are not. And consequentially, the same number of nations in each group (with no reference to how the nations come to be in each group) indicate the two features have equal “weighting”, when this is not necessarily the case. One doesn’t have the full picture of any possible association between the features.
  • Deceptive methods - although this looks unintentional, the use of a thematic map/ choropleth to highlight the mutual exclusivity between emitters (red) and vulnerable nations (blue) is deceptive. It potentially indicates incorrectly that the world has an issue with an abundance of CO2 emitters. Additionally, the lower emitters are not discernible. The may not be the main source of climate change.

Reference

Code

The following code was used to fix the issues identified in the original.

library(readr)
library(dplyr)
library(stringr)
library(ggplot2)
library(extrafont)
#Source data

# Carbon Dioxide Information Analysis Center at Appalachian State University, Boone North Carolina, https://energy.appstate/CDIAC.
c_emissions <- read_csv("nation.1751_2016.csv", skip = 4)

# Notre Dame Global Adaptation Initiative, University of Notre Dame,  https://gain.nd.edu/our-work/country-index/rankings/
gain <- read_csv("resources_2020_04_05_19h09/resources/gain/gain.csv")

# ISO 3166 Countries with Regional Codes, GitHub, https://github.com/lukes/ISO-3166-Countries-with-Regional-Codes/blob/master/all/all.csv
iso_country <- read_csv("iso_country.txt")


c_emissions$`Per capita CO2 emissions (metric tons of carbon)` <- c_emissions$`Per capita CO2 emissions (metric tons of carbon)` %>% parse_double(na= c("."), trim_ws = TRUE)
# Prepare data

# Filter 2016 emissions
c_emi_2016 <- c_emissions %>% filter(Year == 2016)
c_emi_2016 <- c_emi_2016 %>% select(1:3, 9)
names(c_emi_2016)[3] <- "tot_co2"
names(c_emi_2016)[4] <- "capita_co2"


# Clean up join column Nation
c_emi_2016 <- c_emi_2016 %>% 
  mutate(
    Nation = str_replace_all(Nation,  c(" AND " = " & ", " \\(.*\\)" = "", "UNITED STATES OF AMERICA" = "UNITED STATES", "DEMOCRATIC REPUBLIC OF THE CONGO" = "ZAIRE","COTE D IVOIRE" = "COTE D'IVOIRE", "BRUNEI DARUSSALAM"= "BRUNEI", "REPUBLIC OF " = "", "ISLAMIC IRAN" = "IRAN", "UNITED TANZANIA" = "TANZANIA", "LAO PEOPLE S DEMOCRATIC REPUBLIC" = "LAO PEOPLE'S DEMOCRATIC REPUBLIC", "PLURINATIONAL STATE OF BOLIVIA" = "BOLIVIA", "GUINEA BISSAU" = "GUINEA-BISSAU", "RUSSIAN FEDERATION" = "RUSSIA", "LAO PEOPLE'S DEMOCRATIC REPUBLIC" = "LAOS", "UNITED ARAB EMIRATES" = "U A E", "CENTRAL AFRICAN REPUBLIC" = "C A R"))
  )

# Filter 2016 Gains & clean join column
gain_2016 <- gain %>% select(ISO3, Name, `2016`)
gain_2016 <- gain_2016 %>%
              mutate(
                Nation = toupper(Name) %>% str_replace_all(c(" AND " = " & ",", REPUBLIC OF" = "", ", BOLIVARIAN REPUBLIC O" = "", "BRUNEI DARUSSALAM"= "BRUNEI", ", UNITED REPUBLIC OF" = "", ", ISLAMIC REPUBLIC OF" = "", "CONGO, THE DEMOCRATIC REPUBLIC O" = "ZAIRE", ", PLURINATIONAL STATE OF" = "", "RUSSIAN FEDERATION" = "RUSSIA","LAO PEOPLE'S DEMOCRATIC REPUBLIC" = "LAOS", "UNITED ARAB EMIRATES" = "U A E", "CENTRAL AFRICAN REPUBLIC" = "C A R"))
                            
              )
names(gain_2016)[3] <- "gain"


# Join data sets for viz data  
vizdata <- c_emi_2016 %>% inner_join(gain_2016 , by = "Nation") 


#Only use completed cases : where nation has both emission and gain information.
vizdata <- vizdata[complete.cases(vizdata), ]

#Join to get continent info.
vizdata <- vizdata %>% inner_join(iso_country, by = c("ISO3" = "alpha-3"))
#Prepare viz data : factors, ranks for both emissions and gains.  Filter top emitters and most vulnerable.  And categorise them.
vizdata$Nation <- vizdata$Nation %>% factor()
vizdata$ISO3 <- vizdata$ISO3 %>% factor()

vizdata <- vizdata %>% mutate (
                                rnk_capco2 = rank(-capita_co2, ties.method = "min"),
                                rnk_vulnerable = rank(gain, ties.method = "min")
                        )

top_countries <- 30
cat_lbl1 <-  paste(top_countries, "Most vulnerable")
cat_lbl2 <-  paste("Top", top_countries, "CO2 emitters")
cat_lbl3 <-  "Both"


vizdata <- vizdata %>% filter(rnk_capco2 <= top_countries | rnk_vulnerable <= top_countries)

vizdata <- vizdata %>% mutate(
                                cat = case_when(
                                              rnk_vulnerable <= top_countries & rnk_capco2 > top_countries ~ cat_lbl1,
                                              rnk_vulnerable > top_countries & rnk_capco2 <= top_countries ~ cat_lbl2,
                                              TRUE ~ cat_lbl3
                                               
                                )
                              )

vizdata$cat <- vizdata$cat %>% factor(levels = c(cat_lbl2, cat_lbl1, cat_lbl3))
vizdata$region <- vizdata$region %>% factor(levels = c("Africa"  , "Americas", "Asia"   ,  "Europe"  , "Oceania" ))
# configure and plot viz

#colour palette 
pal <- c('#377eb8','#ffbe0b','#e41a1c','#ff7f00','#984ea3')

#background colour
backgrd <- "grey95"

plt <- vizdata %>% ggplot(aes(x = gain, y=capita_co2 ))

plt <- plt +
        geom_point(aes(colour=region), stat = "identity", position = "identity",  size=6, alpha = 0.75) +
        geom_text( aes(label = str_wrap(str_to_title(Nation), width=8)), size=2.25 , hjust = 0.5, vjust = 0.5 , alpha = 0.8, check_overlap = TRUE) +

        facet_grid(rows = vars(cat), scales = "free_y") +
        labs (
          x= "More          Vulnerability to climate change (ND-Gain Index)           Less",
          y = expression("Emitted CO"[2]*" per capita (metric tons of carbon)"),
          title = expression("The Dichotomy : CO"[2]*" emission and climate change vulnerability (2016)"),
          subtitle = "African nations the lowest emitters and most vulnerable. Asian/European nations are the top emitters",
          caption = "Source: CDIAC, Appalachian State University, https://energy.appstate.edu/research/work-areas/cdiac-appstate\nNotre Dame Global Adaptation Initiative, University of Notre Dame,  https://gain.nd.edu/our-work/country-index/rankings"
        ) +
        
        scale_colour_manual( values = pal) +
        theme_gray() +
        theme(
          text=element_text(family="Arial"),
          legend.title = element_blank(),
          plot.background = element_rect(fill=backgrd),
          panel.background = element_rect(fill = backgrd),
          legend.background = element_rect(fill = backgrd),
          panel.grid.minor.x = element_blank(),
          panel.grid.minor.y = element_blank(),
          axis.ticks = element_blank(),
          plot.caption= element_text(size=7.5, hjust=1, margin=margin(t=15)),
          plot.subtitle = element_text(size=10)
        )

Data Reference

Reconstruction

The following plot fixes the main issues in the original.