Using R to explore and visualize decoupling

Using dplyr, ggplot2 and tidyr with data from Our World in Data

R
ecology
economics
Author

Mark Isken

Published

February 20, 2025

Decoupling is the notion that economic growth can continue without accompanying increases in energy use, greenhouse gas emissions, resource use, biodiversity loss, pollution and other adverse environmental and sociological impacts. This phenomena is highlighted and relied upon as a critical linchpin in the green growth movement. Without decoupling, it is hard to see how economic growth can continue indefinitely on a planet with finite resources and a finite capacity to absorb the byproducts of perpetual growth.

In this post we are going to use R to explore decoupling and hopefully learn a few ggplot things along the way. The data we’ll use is from the well known Our World in Data (OWID) site which has done a great job of making well curated and carefully processed data about our world available for all of us. Their lead researcher is data scientist Hannah Ritchie, who recently wrote a book entitled “Not the End of the World: How We Can Be the First Generation to Build a Sustainable Planet” (Ritchie (2024)) which features numerous visualizations based on OWID data.

We’ll need a few libraries.

library(dplyr)
library(tidyr)
library(ggplot2)

Creating the dataset

In the decoupling conversation, GDP is generally used as the measure of economic growth - one of the things that is coupled. It is the driving variable. GDP is tricky to measure and not necessarily a great measure of prosperity. For the driven variable, many possibilities exist but essentially it boils down to resource use or its impact (Vadén et al. (2020)). So, decoupling could occur, or not occur, for any subset of potential driven variables and might involve resource use or environmental impact. It also may be temporary or permanent. Its magnitude may be sufficient to avert some pretty bad outcomes or it may not be. Decoupling might be local, regional or global. Of course, the earth probably isn’t interested in local decoupling unless the ideas underpinning it lead to global decoupling.

I created two dataframes, one at the country level and one at the global level containing a range of variables that would seem to be relevant and for which data was generally available from 1980 until the near present. Data acquisition and processing scripts were written using R and can be found in the data_prep.qmd file in the GitHub repository for this post. The variables included in the global dataframe include:

The country level dataframe includes a country identifier. The following aggregate country related fields were populated using the countrycode package.

  • continent - each country belongs to a continent.
  • region23 - each country is classified as being part of one of twenty-three regions
  • iso3c - a unique 3 character code for each country

There are a few computed fields:

  • primary_energy_consumption_Twh - product of population and primary_energy_consumption_per_capita__kwh and rescaled to terrawatt hours.
  • annual_emissions_ghg_total_co2eq - product of population and annual_emissions_ghg_total_co2eq_per_capita.
  • annual_emissions_ghg_fossil_co2eq - product of population and annual_emissions_ghg_fossil_co2eq_per_capita.
  • gdp_total - product of population and gdp_per_capita.

The basic data processing steps were:

  • download a number of different datasets using the Our World in Data API into individual dataframes,
  • combine all the dataframes by left joining on Entity and Year,
  • use dplyr::filter() to only keep data from 1980 on,
  • use dplyr::filter() along with the grepl() function to get rid of rows that are not associated with countries (e.g. continents),
  • create a separate global level dataframe (filtering on `Entity == ‘World’),
  • add variables to the global dataframe that aren’t relevant at the country level,
  • create a few computed features,
  • added codes such as region23 and iso3c,
  • exported the dataframes to CSV files.

The final result is two CSV files, country_metrics.csv and world_metrics.csv. Let’s read these in and take a look.

country_metrics_df <- read.csv('data/owid/country_metrics.csv')
world_metrics_df <- read.csv('data/owid/world_metrics.csv')
summary(world_metrics_df)
    Entity               year        population       
 Length:44          Min.   :1980   Min.   :4.448e+09  
 Class :character   1st Qu.:1991   1st Qu.:5.396e+09  
 Mode  :character   Median :2002   Median :6.296e+09  
                    Mean   :2002   Mean   :6.298e+09  
                    3rd Qu.:2012   3rd Qu.:7.224e+09  
                    Max.   :2023   Max.   :8.092e+09  
                                                      
 primary_energy_consumption_per_capita__kwh
 Min.   :16753                             
 1st Qu.:17641                             
 Median :18067                             
 Mean   :18891                             
 3rd Qu.:20448                             
 Max.   :21394                             
                                           
 annual_emissions_ghg_total_co2eq_per_capita
 Min.   :6.440                              
 1st Qu.:6.753                              
 Median :6.907                              
 Mean   :6.904                              
 3rd Qu.:7.093                              
 Max.   :7.498                              
                                            
 annual_emissions_ghg_fossil_co2eq_per_capita gas_consumption_twh
 Min.   :4.817                                Min.   :14237      
 1st Qu.:4.961                                1st Qu.:19850      
 Median :5.160                                Median :24673      
 Mean   :5.205                                Mean   :26128      
 3rd Qu.:5.454                                3rd Qu.:33333      
 Max.   :5.652                                Max.   :40239      
                                                                 
 coal_consumption_twh oil_consumption_twh tot_fossil_fuel_consumption_twh
 Min.   :20878        Min.   :33100       Min.   : 69136                 
 1st Qu.:25766        1st Qu.:37668       1st Qu.: 83248                 
 Median :28416        Median :43508       Median : 96597                 
 Mean   :33040        Mean   :43571       Mean   :102739                 
 3rd Qu.:42851        3rd Qu.:48850       3rd Qu.:126842                 
 Max.   :45565        Max.   :54564       Max.   :140231                 
                                                                         
 production.Cobalt.Mine.tonnes production.Copper.Mine.tonnes
 Min.   : 18000                Min.   : 7200000             
 1st Qu.: 35700                1st Qu.: 9297500             
 Median : 51800                Median :13650000             
 Mean   : 73541                Mean   :13704773             
 3rd Qu.:104750                3rd Qu.:17125000             
 Max.   :230000                Max.   :22000000             
                                                            
 production.Lithium.Mine.tonnes production.Nickel.Mine.tonnes
 Min.   : 13100                 Min.   : 621000              
 1st Qu.: 19650                 1st Qu.: 968500              
 Median : 30650                 Median :1350000              
 Mean   : 48612                 Mean   :1540250              
 3rd Qu.: 77925                 3rd Qu.:2115000              
 Max.   :180000                 Max.   :3600000              
 NA's   :20                                                  
 production.Steel.Processing..crude.tonnes freshwater_withdrawals_m3
 Min.   :6.440e+08                         Min.   :3.856e+12        
 1st Qu.:7.362e+08                         1st Qu.:3.875e+12        
 Median :8.795e+08                         Median :3.900e+12        
 Mean   :1.127e+09                         Mean   :3.898e+12        
 3rd Qu.:1.542e+09                         3rd Qu.:3.920e+12        
 Max.   :1.950e+09                         Max.   :3.949e+12        
                                           NA's   :36               
 mean_surface_temp_2m under_five_mortality children_per_woman life_expectancy
 Min.   :13.85        Min.   :3.712        Min.   :2.251      Min.   :60.50  
 1st Qu.:14.10        1st Qu.:4.425        1st Qu.:2.584      1st Qu.:64.04  
 Median :14.29        Median :6.032        Median :2.699      Median :66.91  
 Mean   :14.31        Mean   :6.287        Mean   :2.865      Mean   :67.11  
 3rd Qu.:14.47        3rd Qu.:8.143        3rd Qu.:3.176      3rd Qu.:70.83  
 Max.   :14.97        Max.   :9.303        Max.   :3.744      Max.   :73.17  
                      NA's   :11                                             
 gdp_per_capita  ocean_heat_content_noaa_2000m ocean_heat_content_mri_2000m
 Min.   : 5897   Min.   :-2.249                Min.   :-2.8327             
 1st Qu.: 6791   1st Qu.: 1.666                1st Qu.: 0.9673             
 Median : 7953   Median : 8.016                Median : 7.5673             
 Mean   : 8263   Mean   : 9.976                Mean   :10.2786             
 3rd Qu.: 9651   3rd Qu.:16.803                3rd Qu.:17.0373             
 Max.   :11579   Max.   :29.105                Max.   :29.2773             
                                               NA's   :3                   
 ocean_heat_content_iap_2000m primary_energy_consumption_Twh
 Min.   :-4.85103             Min.   : 77357                
 1st Qu.: 0.02597             1st Qu.: 96337                
 Median : 8.30797             Median :113072                
 Mean   :10.20616             Mean   :120461                
 3rd Qu.:18.53997             3rd Qu.:148203                
 Max.   :29.42097             Max.   :173112                
 NA's   :3                                                  
 annual_emissions_ghg_total_co2eq annual_emissions_ghg_fossil_co2eq
 Min.   :3.232e+10                Min.   :2.302e+10                
 1st Qu.:3.733e+10                1st Qu.:2.722e+10                
 Median :4.150e+10                Median :3.070e+10                
 Mean   :4.337e+10                Mean   :3.298e+10                
 3rd Qu.:5.116e+10                3rd Qu.:4.078e+10                
 Max.   :5.382e+10                Max.   :4.411e+10                
                                                                   
   gdp_total        
 Min.   :2.657e+13  
 1st Qu.:3.658e+13  
 Median :5.007e+13  
 Mean   :5.389e+13  
 3rd Qu.:6.972e+13  
 Max.   :9.369e+13

The country_metrics_df dataframe has most of the same columns but is by country by year.

Defining decoupling is tricky

Decoupling is a complex topic. I highly recommend Parrique et al. (2019) to get a concise overview of what decoupling is and analysis of whether or not or to what degree it is happening and is likely to happen in the future.

For this post, a couple of concepts are important. Relative decoupling is when the driven variable is growing slower than the driving variable (i.e. GDP). Absolute decoupling is when the driving variable is not growing (decreasing hopefully) even though the driving variable continues to grow. In the EEB study, it is argued that:

The validity of the green growth discourse relies on the assumption of an absolute, permanent, global, large and fast enough decoupling of economic growth from all critical environmental pressures.

There’s a lot to chew on there and we certainly aren’t going to get into all of it in this post. Instead, this is just a glimpse at a handful of metrics to see if there’s evidence of decoupling and if it’s relative or absolute. We’ll use R to do it and hopefully learn a few dplyr, tidyr, and ggplot2 tricks along the way.

Basic time series plots of GDP vs primary energy consumption

What if we want to compare two time series such as gdp_total and primary_energy_consumption_Twh? The units are different so we really shouldn’t plot together with two y-axes. We could facet them using a long version of the data.

world_metrics_df |> 
  select(year, gdp_total, primary_energy_consumption_Twh) |> 
  pivot_longer(cols = c(gdp_total, primary_energy_consumption_Twh), names_to = 'metric', values_to = 'value') |> 
  ggplot(aes(x = year, y = value)) +
  geom_line() +
  facet_grid(metric ~ ., scales = 'free_y')

The slopes look similar but that’s just due to the axis limits on the graphs. We could also scatter them against each other.

world_metrics_df |> 
  ggplot() +
  geom_point(aes(x = gdp_total, y = primary_energy_consumption_Twh, colour = year))

Looks like GDP and PEC (primary energy consumption) are not decoupled.

Another option is to index the different series to year 1. Since world_metrics_df is ordered by year, this is relatively straightforward to do for that dataframe. However, for country_metrics_df this will require some careful dplyr use.

For world_metrics_df we can use mutate() to compute the indexed series. For this first plot, I’ll use separate geom_line() objects. The downside of doing plots in this way is that it’s tricky to create a legend. You have to map color in aes() to a string representing the variable as you want it to read in the legend and then use scale_color_manual with a named vector of colors. See https://forum.posit.co/t/adding-manual-legend-to-ggplot2/41651. That is not straight forward and definitely not the “tidy” way to do it.

colors <- c("gdp_total_idx" = "blue", "primary_energy_consumption_Twh_idx" = "black")

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  ggplot() +
  geom_line(aes(x=year, y = gdp_total_idx, color = "gdp_total_idx"), linetype = 'solid') +
  geom_line(aes(x=year, y = primary_energy_consumption_Twh_idx, 
                color = "primary_energy_consumption_Twh_idx"), linetype = 'dashed') +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       color = "Legend", title = "Primary energy consumption") +
    scale_color_manual(values = colors)

From this plot, we can see that the two series exhibit relative decoupling. The overall rate of increase for PEC is less than that of GDP, but it’s still increasing.

Instead of two geoms, we could pivot to long data and filter and facet by metric. This gives us the automatic legend creation and is the tidy approach.

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  select(year, gdp_total_idx, primary_energy_consumption_Twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, primary_energy_consumption_Twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Primary Energy Consumption")

Let’s take a closer look at the most recent few years.

world_metrics_df |> 
  filter(year >= 2020) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  select(year, gdp_total_idx, primary_energy_consumption_Twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, primary_energy_consumption_Twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1", 
       title = "Primary Energy Consumption")

What if we use per-capita values for these same two metrics?

world_metrics_df |> 
  mutate(gdp_per_capita_idx = gdp_per_capita / gdp_per_capita[1],
         primary_energy_consumption_per_capita__kwh_idx = primary_energy_consumption_per_capita__kwh / primary_energy_consumption_per_capita__kwh[1]) |> 
  select(year, gdp_per_capita_idx, primary_energy_consumption_per_capita__kwh_idx) |> 
  pivot_longer(cols = c(gdp_per_capita_idx, primary_energy_consumption_per_capita__kwh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Primary Energy Consumption (per capita)")

So, one sees a greater degree of relative decoupling in the per capita versions of these two metrics. From the earth’s perspective, it doesn’t really care so much about per capita statistics - it’s the overall amount that really matters.

Looking at the last few years, the slopes of the two series are more similar than in the plot starting at 1980. When looking at decoupling, the date range can matter. One might get relative (or even absolute) decoupling for a period of time, only to see that revert to more coupled behavior.

world_metrics_df |> 
  filter(year >= 2020) |>
  mutate(gdp_per_capita_idx = gdp_per_capita / gdp_per_capita[1],
         primary_energy_consumption_per_capita__kwh_idx = primary_energy_consumption_per_capita__kwh / primary_energy_consumption_per_capita__kwh[1]) |> 
  select(year, gdp_per_capita_idx, primary_energy_consumption_per_capita__kwh_idx) |> 
  pivot_longer(cols = c(gdp_per_capita_idx, primary_energy_consumption_per_capita__kwh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Primary Energy Consumption (per capita)")

Greenhouse gas emissions

There was a high profile study done by the World Resources Institute in 2016 entitled The Roads to Decoupling: 21 Countries Are Reducing Carbon Emissions While Growing GDP. Reading the report, it would seem the analysis was production based and not consumption based - a comment is made near the end:

Beyond the aggregate trends described here, more information is needed on the potential leakage of carbon emissions to other countries as nations move their industries overseas, factors that enable sustained and absolute decoupling, and what’s needed to support larger-scale emissions mitigation.

Greenhouse gas emissions are measured in tonnes per person of carbon dioxide-equivalents over a 100-year timescale. See https://ourworldindata.org/grapher/total-greenhouse-gas-emissions-per-capita and https://ourworldindata.org/grapher/per-capita-ghg-excl-land-use.

Let’s start by looking at the whole world and just at fossil fuel emissions.

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         annual_emissions_ghg_fossil_co2eq_idx = annual_emissions_ghg_fossil_co2eq / annual_emissions_ghg_fossil_co2eq[1]) |> 
  select(year, gdp_total_idx, annual_emissions_ghg_fossil_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_fossil_co2eq_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (excluding land use)")

This plot reveals relative decoupling at the global level but still we have emissions increasing. That’s not terribly surprising if one keeps up with the various attempts at forging international agreements for reducing greenhouse gas emissions. There have been decades of talk with not much in terms of results. Meanwhile, the world keeps warming and the outlook isn’t great for our climate.

The last few years look like this.

world_metrics_df |> 
  filter(year >= 2020) |>
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         annual_emissions_ghg_fossil_co2eq_idx = annual_emissions_ghg_fossil_co2eq / annual_emissions_ghg_fossil_co2eq[1]) |> 
  select(year, gdp_total_idx, annual_emissions_ghg_fossil_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_fossil_co2eq_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF"))+
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (excluding land use)")

Now let’s consider greenhouse gas emissions that also factor in land use changes and forestry.

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  select(year, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (including land use)")

world_metrics_df |> 
  filter(year >= 2020) |>
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  select(year, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (including land use)")

Still seeing relative decoupling.

What about those 21 countries in the 2016 report mentioned above?

wri_21 <- c('Austria',
  'Belgium',
  'Bulgaria',
  'Czechia',
  'Denmark',
  'Finland',
  'France',
  'Germany',
  'Hungary',
  'Ireland',
  'Netherlands',
  'Portugal',
  'Romania',
  'Slovakia',
  'Spain',
  'Sweden',
  'Switzerland',
  'Ukraine',
  'United Kingdom',
  'United States',
  'Uzbekistan'
  )

Here we need to be careful when computing the indexed values of our metrics. Notice we use a pattern that looks like this:

  arrange(country, year) |> 
  group_by(country, year) |> 
  mutate(indexed_variable = value / value[1]) |> 
  ungroup()

We need to do a group_by() so that we divide by the group level base value. Then we need to ungroup() to get our detailed level data back for plotting.

country_metrics_df |> 
  filter(country %in% wri_21, year < 2023) |> 
  arrange(country, year) |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  ungroup() |> 
  select(year, country, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  facet_wrap(~country, scales = 'free_y') +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (including land use)")

For these 21 countries, absolute decoupling of overall greenhouse gas emissions from GDP seems to be ongoing for the most part. Of course, since globally we definitely do not see absolute decoupling, there must be several populous countries for which absolute decoupling is not happening. For example:

country_metrics_df |> 
  filter(country %in% c('China', 'India', 'Brazil', 'Mexico', 'South Africa',
                        'Vietnam', 'Venezuela', 'Iran', 'Saudi Arabia')) |> 
  arrange(country, year) |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  ungroup() |> 
  select(year, country, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  facet_wrap(~country, scales = 'free_y') +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Annual GHG emissions (including land use)")
Warning: Removed 1 row containing missing values or values outside the scale range
(`geom_line()`).

Fossil fuel consumption

In countries such as the US, coal plants have been closing in recent years. What does coal consumption look like on a global scale?

world_metrics_df |>
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         coal_consumption_twh_idx = coal_consumption_twh / coal_consumption_twh[1]) |> 
  select(year, gdp_total_idx, coal_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, coal_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Coal consumption")

Looks like coal use is back on the rise.

world_metrics_df |> 
  filter(year >= 2020) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         coal_consumption_twh_idx = coal_consumption_twh / coal_consumption_twh[1]) |> 
  select(year, gdp_total_idx, coal_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, coal_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Coal consumption")

What about gas and oil?

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         oil_consumption_twh_idx = oil_consumption_twh / oil_consumption_twh[1]) |> 
  select(year, gdp_total_idx, oil_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, oil_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Oil consumption")

world_metrics_df |> 
  filter(year >= 2020) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         oil_consumption_twh_idx = oil_consumption_twh / oil_consumption_twh[1]) |> 
  select(year, gdp_total_idx, oil_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, oil_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Oil consumption")

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         gas_consumption_twh_idx = gas_consumption_twh / gas_consumption_twh[1]) |> 
  select(year, gdp_total_idx, gas_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, gas_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Gas consumption")

world_metrics_df |> 
  filter(year >= 2020) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         gas_consumption_twh_idx = gas_consumption_twh / gas_consumption_twh[1]) |> 
  select(year, gdp_total_idx, gas_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, gas_consumption_twh_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Gas consumption")

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         tot_fossil_fuel_consumption_idx = tot_fossil_fuel_consumption_twh / tot_fossil_fuel_consumption_twh[1]) |> 
  select(year, gdp_total_idx, tot_fossil_fuel_consumption_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, tot_fossil_fuel_consumption_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Total fossil fuel consumption")

world_metrics_df |> 
  filter(year >= 2020) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         tot_fossil_fuel_consumption_idx = tot_fossil_fuel_consumption_twh / tot_fossil_fuel_consumption_twh[1]) |> 
  select(year, gdp_total_idx, tot_fossil_fuel_consumption_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, tot_fossil_fuel_consumption_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1",
       title = "Total fossil fuel consumption")

Safe to say there’s no absolute decoupling of fossil fuel use and GDP.

Minerals and metal production

We will look at cobalt, copper, lithium, nickel and steel production.

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         production.Steel.Processing..crude.tonnes_idx = production.Steel.Processing..crude.tonnes / production.Steel.Processing..crude.tonnes[1]) |> 
  select(year, gdp_total_idx, production.Steel.Processing..crude.tonnes_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, production.Steel.Processing..crude.tonnes_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1",
       title = "Steel production")

Let’s look at the sum of cobalt, copper, lithium and nickel production.

world_metrics_df |> 
  mutate(CCLN_production = rowSums(cbind(production.Cobalt.Mine.tonnes,
           production.Copper.Mine.tonnes,
           production.Lithium.Mine.tonnes,
           production.Nickel.Mine.tonnes), na.rm = T)) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         CCLN_production_idx = CCLN_production / CCLN_production[1]) |> 
  select(year, gdp_total_idx, CCLN_production_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, CCLN_production_idx), 
               names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total cobalt, copper, lithium, and nickel production")

No decoupling here.

Water

Water consumption is another important resource use to consider. Data availability only lets us look at 2014-2021.

world_metrics_df |> 
  filter(between(year, 2014, 2021)) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         freshwater_withdrawals_m3_idx = freshwater_withdrawals_m3 / freshwater_withdrawals_m3[1]) |> 
  select(year, gdp_total_idx, freshwater_withdrawals_m3_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, freshwater_withdrawals_m3_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (2014 is year 1)", y = "Multiple of year 1", 
       title = "Freshwater withdrawals")

This will be one to watch as temperatures rise and the boom in data center construction continues to meet the energy needs of the recent advances in AI models. Data centers needs plenty of water for cooling.

US, Canada, Mexico

Let’s look at these three countries in comparison to each other and as a group.

north_america_metrics_df <- country_metrics_df |> 
  filter(country == 'United States' | country == 'Canada' | country == 'Mexico') |> 
  group_by(year) |> 
  summarise(
    tot_fossil_fuel_consumption_twh = sum(tot_fossil_fuel_consumption_twh),
    primary_energy_consumption_Twh = sum(primary_energy_consumption_Twh),
    annual_emissions_ghg_total_co2eq = sum(annual_emissions_ghg_total_co2eq),
    gdp_total = sum(gdp_total)
  )

Primary energy consumption in US, CA, MX

north_america_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  select(year, gdp_total_idx, primary_energy_consumption_Twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, primary_energy_consumption_Twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Primary energy consumption - US, CA, MX")

country_metrics_df |> 
  filter(country == 'United States' | country == 'Canada' | country == 'Mexico') |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  ungroup() |> 
  select(country, year, gdp_total_idx, primary_energy_consumption_Twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, primary_energy_consumption_Twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  facet_grid(country ~ ., scales = 'free_y') +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Primary energy consumption - US, CA, MX")

Total fossil fuel consumption in US, CA, MX

north_america_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       tot_fossil_fuel_consumption_twh_idx = tot_fossil_fuel_consumption_twh / tot_fossil_fuel_consumption_twh[1]) |> 
  select(year, gdp_total_idx, tot_fossil_fuel_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, tot_fossil_fuel_consumption_twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total fossil fuel consumption - US, CA, MX")

country_metrics_df |> 
  filter(country == 'United States' | country == 'Canada' | country == 'Mexico') |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       tot_fossil_fuel_consumption_twh_idx = tot_fossil_fuel_consumption_twh / tot_fossil_fuel_consumption_twh[1]) |> 
  ungroup() |> 
  select(country, year, gdp_total_idx, tot_fossil_fuel_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, tot_fossil_fuel_consumption_twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  facet_grid(country ~ ., scales = 'free_y') +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total fossil fuel consumption - US, CA, MX")

Total greenhouse gas emissions in US, CA, MX

north_america_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  select(year, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total greenhouse gas emissions - US, CA, MX")

This looks promising.

country_metrics_df |> 
  filter(country == 'United States' | country == 'Canada' | country == 'Mexico') |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  ungroup() |> 
  select(country, year, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + facet_grid(country ~ ., scales = 'free_y') +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total greenhouse gas emissions - US, CA, MX")

Recent increases in wind and solar likely the cause of the stabilization of GHG emissions in US and Canada. Mexico is still seeing increases.

What about coal production in these three countries?

country_metrics_df |> 
  filter(country == 'United States' | country == 'Canada' | country == 'Mexico') |> 
  group_by(country) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       coal_consumption_twh_idx = coal_consumption_twh / coal_consumption_twh[1]) |> 
  ungroup() |> 
  select(country, year, gdp_total_idx, coal_consumption_twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, coal_consumption_twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + facet_grid(country ~ ., scales = 'free_y') +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Coal production - US, CA, MX")

Regional summaries

NOTE Missing data series in some of the following plots are due to lack of data in the index year of 1980.

country_metrics_df |> 
  filter(year <= 2022) |> 
  group_by(year, region23) |> 
  summarise(
    tot_fossil_fuel_consumption_twh = sum(tot_fossil_fuel_consumption_twh, na.rm = TRUE),
    primary_energy_consumption_Twh = sum(primary_energy_consumption_Twh, na.rm = TRUE),
    annual_emissions_ghg_total_co2eq = sum(annual_emissions_ghg_total_co2eq, na.rm = TRUE),
    gdp_total = sum(gdp_total, na.rm = TRUE)
  ) |> 
  arrange(region23, year) |> 
  group_by(region23) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       primary_energy_consumption_Twh_idx = primary_energy_consumption_Twh / primary_energy_consumption_Twh[1]) |> 
  ungroup() |> 
  select(year, region23, gdp_total_idx, primary_energy_consumption_Twh_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, primary_energy_consumption_Twh_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  facet_wrap(~region23, scales = 'free_y') +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Primary energy consumption")

country_metrics_df |> 
  filter(year <= 2022) |> 
  group_by(year, region23) |> 
  summarise(
    tot_fossil_fuel_consumption_twh = sum(tot_fossil_fuel_consumption_twh, na.rm = TRUE),
    primary_energy_consumption_Twh = sum(primary_energy_consumption_Twh, na.rm = TRUE),
    annual_emissions_ghg_total_co2eq = sum(annual_emissions_ghg_total_co2eq, na.rm = TRUE),
    gdp_total = sum(gdp_total, na.rm = TRUE)
  ) |> 
  arrange(region23, year) |> 
  group_by(region23) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       annual_emissions_ghg_total_co2eq_idx = annual_emissions_ghg_total_co2eq / annual_emissions_ghg_total_co2eq[1]) |> 
  ungroup() |> 
  select(year, region23, gdp_total_idx, annual_emissions_ghg_total_co2eq_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, annual_emissions_ghg_total_co2eq_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + facet_wrap(~region23, scales = 'free_y') +
  scale_color_manual(values = c("#A3BAC2FF", "#6C5043FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Total greenhouse gas emissions")

Population

Population marches steadily on.

world_metrics_df |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         population_idx = population / population[1]) |> 
  select(year, gdp_total_idx, population_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, population_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1980 is year 1)", y = "Multiple of year 1", 
       title = "Population")

world_metrics_df |> 
  filter(year >= 2020) |>
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
         population_idx = population / population[1]) |> 
  select(year, gdp_total_idx, population_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, population_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + 
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (2020 is year 1)", y = "Multiple of year 1", 
       title = "Population")

country_metrics_df |> 
  filter(between(year, 1990, 2022)) |> 
  group_by(year, region23) |> 
  summarise(
    population = sum(population, na.rm = TRUE),
    gdp_total = sum(gdp_total, na.rm = TRUE)
  ) |> 
  arrange(region23, year) |> 
  group_by(region23) |> 
  mutate(gdp_total_idx = gdp_total / gdp_total[1],
       population_idx = population / population[1]) |> 
  ungroup() |> 
  select(year, region23, gdp_total_idx, population_idx) |> 
  pivot_longer(cols = c(gdp_total_idx, population_idx), names_to = 'metric', values_to = 'value') |> 
  ggplot() +
  geom_line(aes(x=year, y = value, color = metric)) + facet_wrap(~region23, scales = 'free_y') +
  scale_color_manual(values = c("#6C5043FF", "#A3BAC2FF")) +
  labs(x = "Year (1990 is year 1)", y = "Multiple of year 1", title = "Population")

Parting words

From p3 of the 2019 EEB study:

The fact that decoupling on its own, i.e. without addressing the issue of economic growth, has not been and will not be sufficient to reduce environmental pressures to the required extent is not a reason to oppose decoupling (in the literal sense of separating the environmental pressures curve from the GDP curve) or the measures that achieve decoupling - on the contrary, without many such measures the situation would be far worse. It is a reason to have major concerns about the predominant focus of policymakers on green growth, this focus being based on the flawed assumption that sufficient decoupling can be achieved through increased efficiency without limiting economic production and consumption.

Data source citations

Here are the citations from the various OWID data sources used as part of this post.

  • HYDE (2023); Gapminder (2022); UN WPP (2024) – with major processing by Our World in Data. “Population” [dataset]. PBL Netherlands Environmental Assessment Agency, “History Database of the Global Environment 3.3”; Gapminder, “Population v7”; United Nations, “World Population Prospects”; Gapminder, “Systema Globalis” [original data]. Retrieved February 13, 2025 from https://ourworldindata.org/grapher/population
  • Bolt and van Zanden - Maddison Project Database 2023 – with minor processing by Our World in Data. “GDP per capita – Maddison Project Database – In constant international-$. Historical data” [dataset]. Bolt and van Zanden, “Maddison Project Database 2023” [original data]. Retrieved February 13, 2025 from https://ourworldindata.org/grapher/gdp-per-capita-maddison-project-database
  • U.S. Energy Information Administration (2023); Energy Institute - Statistical Review of World Energy (2024); Population based on various sources (2023) – with major processing by Our World in Data. “Primary energy consumption per capita” [dataset]. U.S. Energy Information Administration, “International Energy Data”; Energy Institute, “Statistical Review of World Energy”; Various sources, “Population” [original data]. Retrieved February 17, 2025 from https://ourworldindata.org/grapher/per-capita-energy-use
  • Jones et al. (2024); Population based on various sources (2024) – with major processing by Our World in Data. “Per capita greenhouse gas emissions including land use” [dataset]. Jones et al., “National contributions to climate change 2024.2”; Various sources, “Population” [original data]. Retrieved February 17, 2025 from https://ourworldindata.org/grapher/total-greenhouse-gas-emissions-per-capita
  • Jones et al. (2024); Population based on various sources (2024) – with major processing by Our World in Data. “Per capita greenhouse gas emissions from fossil fuels and industry” [dataset]. Jones et al., “National contributions to climate change 2024.2”; Various sources, “Population” [original data]. Retrieved February 13, 2025 from https://ourworldindata.org/grapher/per-capita-ghg-excl-land-use
  • Energy Institute - Statistical Review of World Energy (2024) – with major processing by Our World in Data. “Coal consumption” [dataset]. Energy Institute, “Statistical Review of World Energy” [original data].
  • USGS - Mineral Commodity Summaries (2024); USGS - Historical Statistics for Mineral and Material Commodities (2023); BGS - World Mineral Statistics (2023) – with major processing by Our World in Data. “Copper production” [dataset]. United States Geological Survey, “Mineral Commodity Summaries”; United States Geological Survey, “Historical Statistics for Mineral and Material Commodities”; British Geological Survey, “World Mineral Statistics” [original data].
  • Food and Agriculture Organization of the United Nations (via World Bank) (2025) – processed by Our World in Data. “Annual freshwater withdrawals” [dataset]. Food and Agriculture Organization of the United Nations (via World Bank), “World Development Indicators” [original data]. Retrieved February 13, 2025 from https://ourworldindata.org/grapher/annual-freshwater-withdrawals
  • Contains modified Copernicus Climate Change Service information (2025) – with major processing by Our World in Data. “Annual average” [dataset]. Contains modified Copernicus Climate Change Service information, “ERA5 monthly averaged data on single levels from 1940 to present 2” [original data].
  • Gapminder (2020); UN Inter-agency Group for Child Mortality Estimation (2024) – with major processing by Our World in Data. “Under-five mortality rate – UN IGME; Gapminder – Long-run data” [dataset]. United Nations Inter-agency Group for Child Mortality Estimation, “United Nations Inter-agency Group for Child Mortality Estimation”; Gapminder, “Child mortality rate under age five v7”; Various sources, “Population” [original data]. Retrieved February 20, 2025 from https://ourworldindata.org/grapher/child-mortality
  • UN, World Population Prospects (2024) – processed by Our World in Data. “Fertility rate, total – UN WPP” [dataset]. United Nations, “World Population Prospects” [original data]. Retrieved February 20, 2025 from https://ourworldindata.org/grapher/children-per-woman-un
  • World Bank Poverty and Inequality Platform (2024) – with major processing by Our World in Data. “Gini Coefficient – World Bank” [dataset]. World Bank Poverty and Inequality Platform, “World Bank Poverty and Inequality Platform (PIP) 20240627_2017, 20240627_2011” [original data].
  • UN WPP (2024); HMD (2024); Zijdeman et al. (2015); Riley (2005) – with minor processing by Our World in Data. “Life expectancy at birth – Various sources – period tables” [dataset]. Human Mortality Database, “Human Mortality Database”; United Nations, “World Population Prospects”; Zijdeman et al., “Life Expectancy at birth 2”; James C. Riley, “Estimates of Regional and Global Life Expectancy, 1800-2001” [original data]. Retrieved February 20, 2025 from https://ourworldindata.org/grapher/life-expectancy
  • U.S. Bureau of Economic Analysis via FRED®

References

Parrique, Timothée, Jonathan Barth, François Briens, Christian Kerschner, Alejo Kraus-Polk, Anna Kuokkanen, and Joachim H Spangenberg. 2019. “Decoupling Debunked.” Evidence and Arguments Against Green Growth as a Sole Strategy for Sustainability. A Study Edited by the European Environment Bureau EEB 3.
Ritchie, Hannah. 2024. Not the End of the World: How We Can Be the First Generation to Build a Sustainable Planet. Random House.
Vadén, Tere, Ville Lähde, Antti Majava, Paavo Järvensivu, Tero Toivanen, Emma Hakala, and Jussi T Eronen. 2020. “Decoupling for Ecological Sustainability: A Categorisation and Review of Research Literature.” Environmental Science & Policy 112: 236–44.

Citation

BibTeX citation:
@online{isken2025,
  author = {Isken, Mark},
  title = {Using {R} to Explore and Visualize Decoupling},
  date = {2025-02-20},
  url = {https://bitsofanalytics.org//posts/searching-for-decoupling/searching_for_decoupling.html},
  langid = {en}
}
For attribution, please cite this work as:
Isken, Mark. 2025. “Using R to Explore and Visualize Decoupling.” February 20, 2025. https://bitsofanalytics.org//posts/searching-for-decoupling/searching_for_decoupling.html.