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

Original


Source: The Lancent Planetary Health, Volume3, Issue 10 (2019).


Objective

The visualisation was taken from a journal article, “The carbon footprint of the Chinese health-care system: an environmentally extended input–output and structural path analysis study” published in The Lancent Planetary Health in 2019.

AIM: The article aimed at examining Greenhouse gas(GHG) emissions from the healthcare sector in China. According to energy data from the National Bureau of Statistics based on the figures reported by various enterprises, despite China having a much smaller carbon footprint in the healthcare sector than other developed nations like the USA and Australia, etc., the carbon emissions per unit of healthcare expenditure are very high. Purchases of goods and services are responsible for 84% of all emissions in medical facilities. The findings of the study by the Natural Science Foundation of China recommended that a nationwide carbon-efficient objective should be implemented in the healthcare industry to reduce the emissions.
The visualization aimed to project carbon emissions by different sectors in China’s medical industry, namely transportation and procurement. The innermost layer shows the two categories and the other two layers further represent subsections of respective sectors labeled with their names and emission(in mega tonnes). The main objective of the graph was to depict a breakdown of carbon footprint in these sectors. .

TARGET AUDIENCE: Government policy makers; healthcare professionals; industry, transport and supply sector; the public; and the vulnerable population.

The visualization chosen had the following three main issues:

  • Representation of numerical values: The donut chart had too many categories to compare.Human eye is not very good at comparing slices that have almost similar areas. Also, pie charts usually represent the participation percentage based on the angles of each slice which was not the case here.
  • Colour scheme: Similar sectors(like ‘electricity and steam’, ‘other sectors’ and ‘other services’) had the same colour in every layer, but the colours did not differentiate the layers mentioned in the description. So it is not too clear that three different sectors are mentioned in the graph.
  • Failure to compare subsections of different layers based on area: Every layer has a different orientation. The innermost layer depicts two sectors which fills the ring completely.However, the second and the outermost layer have some empty space in the rings which creates confusion as to whether or not the area occupied by individual sub-sectors can be considered as their contribution in the respective sector. Also, because the areas have a different value for each sector, it seems impossible to compare sectors from different layers(for subsecttions that are part of both layers). The only way to compare was with emission numbers, so donut chart was not the best choice to depict this data.

Reference

  • Wu, R. (2019) “The carbon footprint of the Chinese health-care system: An environmentally extended input–output and structural path analysis study,” The Lancet Planetary Health, 3(10). Available at: https://doi.org/10.1016/s2542-5196(19)30192-5.

Code

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

#loading required libraries 
library(ggplot2)
library(magrittr)
library(stringr)
#Creating a data frame with values obtained from original visualization 
Emissions <- data.frame(Sectors = c('Procurement', 'Procurement', 'Procurement', 'Procurement', 'Procurement', 'Procurement', 'Procurement', 'Procurement', 'Building and Transport', 'Building and Transport', 'Building and Transport', 'Building and Transport', 'Building and Transport', 'Building and Transport', 'Building and Transport'),
                        Subsection = c('Electricity and Steam', 'Pharmaceuticals', 'Medical Equipment', 'Apparel', 'Wholesale , Retail, Hotels and Catering', 'Outsourced transport', 'Other services', 'Other sectors', 'Agriculture', 'Pharmaceuticals', 'Electricity and Steam', 'Chemicals', 'Non-metal', 'Other services', 'Other sectors'),
                        Emission = c(17, 118, 13, 8, 6, 6, 7, 17, 25, 24, 16, 12, 6, 5, 20))

#wraping the headings for sectors and subsection for the purpose of better visualization
Emissions$Sectors<- str_wrap(Emissions$Sectors, width = 22)
Emissions$Subsection <- str_wrap(Emissions$Subsection, width = 25)

#plotting the graph
p1 <-ggplot(Emissions, aes(x = Subsection, y = Emission,fill=Sectors)) +
  geom_col(colour = 'black') +coord_flip() +facet_wrap(~Sectors)+
  geom_text(aes(label=Emission), nudge_y=28.5)+
  labs(title = "Breakdown of Carbon Footprint of Medical Institution Sector")+
  scale_fill_manual(values = c("pink", "lightblue"), name="Sectors\n[Carbon Footprint]", labels=c("Building and Transport\n[34mt]","Procurement\n[174mt]"))+
  ylab("Emissions\n(in Megatonnes)") +xlab("Subsections") +theme_minimal() + scale_y_continuous(limits = c(0,300))+
  theme(panel.border = element_rect(linetype = "dashed", fill = NA), legend.box.margin = margin(116, 6, 6, 6), legend.text = element_text(size = 8))

Data Reference

  • Wu, R. (2019) “The carbon footprint of the Chinese health-care system: An environmentally extended input–output and structural path analysis study,” The Lancet Planetary Health, 3(10). Available at: https://doi.org/10.1016/s2542-5196(19)30192-5.

Reconstruction

Changes made in the original plot: 1. The numerical values are used as the unit values. Comparison is easier now as every bar starts from the same point. Additionally, every bar is labeled with their emission value to depict exact measurement. 2. Every sector has been given a different colour. It is observed that there are subsections that play a part in more than one sector, but the main focus was to understand the participation of those subsections in individual sectors. 3. With the faceted barcharts, it is easy to compare not only the sectors but also subsections in every sector. This could potentially help with future plans of reducing the carbon footprint and decide what subsections they should be given more attention.

The following plot fixes the main issues in the original: