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

Original


*Source: http://wtfviz.net/post/60203066686/the-spiral-staircase-courtesy-of-janwillemtulp.*


Objective

Explain the objective of the original data visualisation and the targetted audience.

The objective of the original data visualisation was to show the components of a Winning Ted talk or an effective ted talk. Meaning what goes into the deliverance of a high quality ted talk. This included things like an opening joke or using sophisticated visual aids. Shapes and percentages were shown to display which component was more important.

The targeted audience is everyone interested in ted talks and especially those individuals who would like to know what goes into delivering an effective ted talk, and which factors hold more relative importance.

The visualisation chosen had the following three main issues:

The first issue lies in the fact that the visualisation is basically a 3d pie chart. This uses volume and angles to display numeric values. According to Cleveland and Mcgill, both of these have low accuracy for the representation of numeric values thus this is an issue

The second issue lies in the use of colour in the visualisation. Very bright and saturated colours are used which strains the eyes and seems artificial.

The third issue lies in the fact that too many colours are used in the visualisation which confuses the viewer. Often less is more and in this visualisation 7 colours are used which overwhelms the senses.

Reference

Skau, D. (2013). Visually Blog Learning From Mistakes in Visualization | Visually Blog. Retrieved 26 September 2019, from https://visual.ly/blog/learning-from-mistakes-in-visualization/?fbclid=IwAR25na8bpIhCMiwB9mIze-wuWwOxW-1EmtoNX5CvkggTIdWxEGAXM5KKLCA

Code

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

library(ggplot2)
library(readxl)
Book1_ass <- read_excel("C:/Users/SAP/Desktop/Book1-ass.xlsx")

x<-ggplot(Book1_ass, aes(x= `Anatomy of a winning Ted Talk`, y=Percentage)) 
x+geom_bar(stat = "identity", width = 0.5, fill="dodgerblue3")+theme_minimal()+theme(axis.text.x = element_text(angle = 45,hjust = 1))+
  labs(title = "Anatomy of a Winning ted talk",
       y="Percentage(%)",
       x="Components" ) + geom_text(aes(label=Percentage), vjust = -0.25,size = 3,colour='black')

Data Reference Skau, D. (2013). Visually Blog Learning From Mistakes in Visualization | Visually Blog. Retrieved 26 September 2019, from https://visual.ly/blog/learning-from-mistakes-in-visualization/?fbclid=IwAR25na8bpIhCMiwB9mIze-wuWwOxW-1EmtoNX5CvkggTIdWxEGAXM5KKLCA

Reconstruction

The following plot fixes the main issues in the original.

I used the original visualisation to create an excel file of percentages and corresponding components which I imported to R. After that I converted into a bar chart using the ggplot function. A bar chart uses a common scale and length to represent numeric values thus I foudn it to be a far better way to represent data than a pie chart. Width of 0.5 was used to have some seperation between the bars so each bar can be seen clearly. I choose the ‘dodgerblue3’ colour to fill the bars as it has low saturation and isnt too bright. A simple white background is used to show a good contrast and attract the viewrs attention towards the blue bars. Above each bar is a black number representing the percentage so the relative difference in percentages between bars can be easily seen. On the x axis the components have been tilted so they can be easily read and dont overlap. On the y axis i choose a scale which shows an accurate visual comparison between values.