Data stories: Climate change

climate change
data science
Published

November 1, 2024

It’s impossible to ignore the big decisions that Americans will be making next Tuesday. But neither campaign is talking about climate change in a serious way. I have my own thoughts about why, but let’s look at some data rather than pontificate.

I found that there is an Emissions Database for Global Atmospheric Research (EDGAR) site that collects data on these questions. Here is the site for emissions data from 1970 to 2024 for a large set of countries:

https://data.jrc.ec.europa.eu/dataset/74f34d14-5819-4954-8b61-b11ce3d4ca48

The following code downloads the data.

url <- "https://jeodpp.jrc.ec.europa.eu/ftp/jrc-opendata/EDGAR/datasets/EDGAR_2024_report/EDGAR_2024_GHG_booklet_2024_fossilCO2only.xlsx"

download.file(url, destfile = "data/edgar-co2-2024.xlsx", mode = "wb")

This is an Excel spreadsheet, so the readxl package will be helpful.

Some poking around in the Excel spreadsheet indicates two tabs that are especially relevant fossil_CO2_per_GDP_by_country and fossil_CO2_per_capita_by_country.

sheets_avail <- readxl::excel_sheets("data/edgar-co2-2024.xlsx")
sheets_avail
[1] "info"                            "citations and references"       
[3] "fossil_CO2_totals_by_country"    "fossil_CO2_by_sector_country_su"
[5] "fossil_CO2_per_GDP_by_country"   "LULUCF_macroregions"            
[7] "fossil_CO2_per_capita_by_countr"
CO2_per_capita <- readxl::read_excel("data/edgar-co2-2024.xlsx", sheet = sheets_avail[7])
str(CO2_per_capita)
tibble [212 × 57] (S3: tbl_df/tbl/data.frame)
 $ Substance         : chr [1:212] "CO2" "CO2" "CO2" "CO2" ...
 $ EDGAR Country Code: chr [1:212] "ABW" "AFG" "AGO" "AIA" ...
 $ Country           : chr [1:212] "Aruba" "Afghanistan" "Angola" "Anguilla" ...
 $ 1970              : num [1:212] 0.427 0.156 1.318 0.34 2.253 ...
 $ 1971              : num [1:212] 0.485 0.152 1.23 0.338 2.197 ...
 $ 1972              : num [1:212] 0.66 0.144 1.461 0.35 2.448 ...
 $ 1973              : num [1:212] 0.735 0.144 1.559 0.323 2.147 ...
 $ 1974              : num [1:212] 0.718 0.178 1.58 0.358 2.26 ...
 $ 1975              : num [1:212] 0.946 0.161 1.419 0.391 2.25 ...
 $ 1976              : num [1:212] 0.931 0.147 0.923 0.367 2.352 ...
 $ 1977              : num [1:212] 1.112 0.175 1.48 0.381 2.463 ...
 $ 1978              : num [1:212] 1.197 0.146 1.697 0.434 2.652 ...
 $ 1979              : num [1:212] 1.266 0.155 1.641 0.48 3.022 ...
 $ 1980              : num [1:212] 1.327 0.151 1.603 0.66 3.011 ...
 $ 1981              : num [1:212] 1.369 0.173 1.431 1 2.586 ...
 $ 1982              : num [1:212] 1.334 0.181 1.333 1.067 2.608 ...
 $ 1983              : num [1:212] 1.357 0.242 1.331 0.933 2.741 ...
 $ 1984              : num [1:212] 1.569 0.268 1.3 0.887 2.906 ...
 $ 1985              : num [1:212] 1.652 0.35 1.329 0.552 2.721 ...
 $ 1986              : num [1:212] 1.82 0.34 1.3 0.46 2.68 ...
 $ 1987              : num [1:212] 2.081 0.285 1.293 0.853 2.657 ...
 $ 1988              : num [1:212] 2.339 0.268 1.298 0.767 2.565 ...
 $ 1989              : num [1:212] 3.235 0.246 1.277 0.773 2.43 ...
 $ 1990              : num [1:212] 3.232 0.239 0.923 0.735 2.025 ...
 $ 1991              : num [1:212] 3.501 0.219 0.938 0.876 1.337 ...
 $ 1992              : num [1:212] 3.672 0.13 0.935 0.881 0.729 ...
 $ 1993              : num [1:212] 3.467 0.118 0.906 1.076 0.697 ...
 $ 1994              : num [1:212] 3.724 0.104 0.83 1.378 0.746 ...
 $ 1995              : num [1:212] 3.9115 0.0856 0.9078 1.6741 0.6663 ...
 $ 1996              : num [1:212] 2.4827 0.0795 1.0679 1.6134 0.6645 ...
 $ 1997              : num [1:212] 4.011 0.0753 1.087 1.3179 0.511 ...
 $ 1998              : num [1:212] 4.0207 0.0753 1.0988 1.1817 0.6169 ...
 $ 1999              : num [1:212] 3.9943 0.0709 1.1305 1.0902 0.9907 ...
 $ 2000              : num [1:212] 2.9561 0.0506 1.0043 1.4045 1.0354 ...
 $ 2001              : num [1:212] 2.9222 0.0451 0.9534 1.4005 1.1015 ...
 $ 2002              : num [1:212] 3.122 0.0429 0.9057 0.9302 1.2845 ...
 $ 2003              : num [1:212] 3.5244 0.0437 0.952 1.1749 1.3536 ...
 $ 2004              : num [1:212] 3.534 0.037 0.938 1.093 1.425 ...
 $ 2005              : num [1:212] 3.8472 0.0505 0.7998 1.0824 1.3473 ...
 $ 2006              : num [1:212] 4.2019 0.0544 0.8054 0.9944 1.3537 ...
 $ 2007              : num [1:212] 4.5202 0.0755 0.8123 0.941 1.4036 ...
 $ 2008              : num [1:212] 4.479 0.14 0.879 1.144 1.402 ...
 $ 2009              : num [1:212] 4.857 0.217 0.935 1.488 1.443 ...
 $ 2010              : num [1:212] 4.743 0.27 0.973 1.745 1.55 ...
 $ 2011              : num [1:212] 2.843 0.384 0.977 1.866 1.704 ...
 $ 2012              : num [1:212] 3.654 0.318 0.985 1.843 1.552 ...
 $ 2013              : num [1:212] 3.996 0.262 1.098 1.955 1.634 ...
 $ 2014              : num [1:212] 4.246 0.239 1.147 1.931 1.745 ...
 $ 2015              : num [1:212] 4.428 0.247 1.188 1.918 1.664 ...
 $ 2016              : num [1:212] 4.626 0.217 1.086 1.921 1.548 ...
 $ 2017              : num [1:212] 4.433 0.227 0.938 1.951 1.823 ...
 $ 2018              : num [1:212] 4.409 0.218 0.853 1.878 1.808 ...
 $ 2019              : num [1:212] 5.261 0.195 0.867 1.819 1.686 ...
 $ 2020              : num [1:212] 4.252 0.185 0.631 1.492 1.55 ...
 $ 2021              : num [1:212] 4.679 0.204 0.745 1.468 1.712 ...
 $ 2022              : num [1:212] 4.698 0.208 0.782 1.457 1.564 ...
 $ 2023              : num [1:212] 4.908 0.214 0.782 1.435 1.558 ...

We need to make this longer if we want to use Tidyverse-style plotting commands.

CO2_per_capita_long <- CO2_per_capita |>
  tidyr::pivot_longer(
    cols = (starts_with("19") |
              starts_with("20")),
    names_to = "year",
    values_to = "emissionsPerCapita"
  ) |>
  # Make the year a properly formatted YYYY-MM-DD string for parsing
  dplyr::mutate(year = lubridate::as_date(paste0(year, "-12-31")))

Now we can make our first attempt at a plot.

library(ggplot2)
CO2_per_capita_long |>
  dplyr::filter(
    Country %in% c(
      "United States",
      "China",
      "Japan",
      "Germany",
      "France and Monaco",
      "Canada",
      "United Kingdom",
      "Taiwan",
      "South Korea",
      "Australia"
    )
  ) |>
  ggplot2::ggplot() +
  aes(year, emissionsPerCapita, color = Country, group = Country) +
  geom_point() +
  geom_line() +
  theme_classic() +
  theme(axis.text.x = element_text(angle = 90))
Figure 1: CO2 emissions per capita 1970-2023 from the Emissions Database for Global Atmospheric Research (EDGAR) database.

Let’s look at the top countries in 2023.

top_2023 <- CO2_per_capita_long |>
  dplyr::filter(year == "2023-12-31") |>
  dplyr::arrange(desc(emissionsPerCapita))

top_20_2023 <- top_2023[1:20,]
top_20_countries_2023 <- top_20_2023$Country

top_20_2023 |>
  dplyr::select(Country, emissionsPerCapita) |>
  kableExtra::kable(format = 'html') |>
  kableExtra::kable_classic()
Country emissionsPerCapita
Palau 62.59089
Qatar 43.55052
Kuwait 24.89597
Brunei 21.12187
New Caledonia 20.90426
Bahrain 20.69999
United Arab Emirates 20.22100
Trinidad and Tobago 19.70759
Gibraltar 19.67394
Saudi Arabia 17.15496
Oman 17.11290
Canada 14.91330
Curaçao 14.66348
Russia 14.45143
Australia 14.21297
United States 13.83075
Seychelles 12.75529
Kazakhstan 12.43028
Taiwan 11.67587
Luxembourg 11.17618

Let’s see how these countries fared since 1970.

CO2_per_capita_long |>
  dplyr::filter(Country %in% top_20_2023$Country) |>
  ggplot2::ggplot() +
  aes(year, emissionsPerCapita, color = Country, group = Country) +
  geom_point() +
  geom_line() +
  theme_classic() +
  theme(axis.text.x = element_text(angle = 90))
Figure 2: CO2 emissions per capita 1970-2023 from the Emissions Database for Global Atmospheric Research (EDGAR) database. This chart focuses on the top 20 emitters in 2023.

Now, let’s see how the top 20 in 1970 have changed.

top_1970 <- CO2_per_capita_long |>
  dplyr::filter(year == "1970-12-31") |>
  dplyr::arrange(desc(emissionsPerCapita))

top_20_1970 <- top_1970[1:20,]
top_20_countries_1970 <- top_20_1970$Country

top_20_1970 |>
  dplyr::select(Country, emissionsPerCapita) |>
  kableExtra::kable(format = 'html') |>
  kableExtra::kable_classic()
Country emissionsPerCapita
Palau 173.76140
Qatar 134.23816
Curaçao 101.06091
United Arab Emirates 82.65335
Luxembourg 57.93929
Kuwait 51.73405
United States 21.92427
Czechia 19.17800
Libya 18.54289
Canada 16.69381
Estonia 15.63304
Brunei 15.62203
Belgium 14.45745
Kazakhstan 14.11505
Bahrain 14.00774
Germany 13.80446
Denmark 12.56345
Australia 12.48760
United Kingdom 12.10428
Sweden 11.93955 
CO2_per_capita_long |>
  dplyr::filter(Country %in% top_20_1970$Country) |>
  ggplot2::ggplot() +
  aes(year, emissionsPerCapita, color = Country, group = Country) +
  geom_point() +
  geom_line() +
  theme_classic() +
  theme(axis.text.x = element_text(angle = 90))
Figure 3: CO2 emissions per capita 1970-2023 from the Emissions Database for Global Atmospheric Research (EDGAR) database. This chart focuses on the top 20 emitters in 1970.

So, the top emitters per capita in 1970 have made substantial progress in reducing emissions. The top emitters in 2023 also made substantial reductions from 1970 to 2023.

The East Asian countries that have have increased emissions per capita since 1970 (China, Taiwan, Japan, South Korea) have joined the mid-ranks of the other wealthy countries in emissions per capita.

This all strikes me as very interesting, and largely positive news. The data story I would tell is that the world is getting richer, and that per capita CO2 emissions are falling into a similar range, even among countries that have grown quickly over the past 50 years. We just need to do more, especially those of us living in wealthy, large countries who can.