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

Original


Source: StatCounter (2020).


Objective

The initial objective of this data visualisation was to provide an insight into the usage of the most common search engines across desktop, tablet and mobile within the top 15 countries in the world in terms of GDP as of 2020. While doing so, the data visualisation in accompaniment with its article also intends to help give an idea of the state of the market of search engines across the world. Giving businesses who are looking for growth some assistance in understanding which search engines suit their needs and will provide their website with as much traffic as possible.

However, the data visualisation produced does have a few fallbacks and issues which need attention and they include:

  • Issue 1 – Data Integrity: right off the bat, it is not immediately clear when exactly this data is sourced from. The data visualisation states that the Source is “StatCounter, 2019”. The title of the accompanying article is called “Global Search Engine Market Share in the Top 15 GDP Nations (Updated for 2020)”. But it’s not until actually looking into the datasets from StatCounter, it can be found that the data is sourced from February 2020. Additionally, some but not all of the data points in the visualisation don’t correspond to that of what is found within the dataset.

  • Issue 2 – Deceptive Methods: although probably not intentional the colouring in of the countries on the map don’t seem to be well thought out and fairly inaccurate evident through how New Zealand appears to be coloured in, in connection with Australia’s statistics. Not to mention the colouring in on this map could be perceived as specific location implies search engine usage. Take Australia for example, it could be viewed as citizens of Melbourne use Google, while citizens of Darwin use Yahoo!, when this is not the case. As such the use of a map in this visualisation highlighting only 15 countries is a questionable one and there are other visualisation types that would be much better suited to this data.

  • Issue 3 – Perceptual/Colour Issues: The last distinct issue is prevalence of blue everywhere within the visualisation. Across the map, the different shadings of blue for each search engine, the blue of the ocean in the map, the blue of the countries boxes and the titles and subtitles. Without some contrasting colours, being able to distinctly tell apart what is what within the visualisation could be a problem, especially if the visualisation is printed in greyscale or if someone who is looking at it has blue colourblind deficiencies.

Reference

Code

The following code was used to fix the issues which were identified in the original data visualisation.

library(ggplot2)

brands <- c('Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Baidu',    'Sogou',    'Shenma',   'Google', 'Other',
            'Google',   'Bing', 'Ecosia',   'Yahoo!', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Bing', 'Yahoo!',   'Ecosia', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'Yahoo!',   'Bing', 'Baidu', 'Other',
            'Google',   'Naver', 'Bing',    'Yahoo!', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other',
            'Google',   'YANDEX RU',    'Mail.ru',  'Bing', 'Other',
            'Google',   'Bing', 'Yahoo!',   'DuckDuckGo', 'Other')
            
countries <- c(rep('Australia', 5), rep('Brazil', 5),   rep('Canada', 5),
               rep('China', 5), rep('Germany', 5), rep('Spain', 5),
               rep('France', 5), rep('UK', 5),  rep('India', 5),    rep('Italy', 5),
               rep('Japan', 5), rep('South Korea', 5),  rep('Mexico', 5),
               rep('Russia', 5), rep('USA', 5))

values <- c(94.69,  3.5,    0.78,   0.63, 0.4,
            97.41,  1.28,   1.11,   0.11, 0.09,
            92.21,  4.54,   2.03,   0.86, 0.36,
            69.12,  20.26,  4.62,   1.63, 4.37,
            94.44,  3,  0.66,   0.59, 1.31,
            96.58,  2.07,   0.8, 0.25, 0.3,
            92.35,  3.66,   1.41,   1.15, 1.43,
            93.01,  4.43,   1.45,   0.58, 0.53,
            98.73,  0.74,   0.42,   0.08, 0.03,
            96.57,  2.04,   0.8,    0.23, 0.36,
            75.09,  18.79,  5.63,   0.2, 0.29,
            74.84,  19.07,  2.08,   1.85, 2.16,
            96.69,  2.06,   0.88,   0.15, 0.22,
            52.29,  44.7,   2.06,   0.36, 0.59,
            88.24,  6.45,   3.61,   1.33, 0.37)

searchengine <- data.frame(brands, countries, values)

fixed <- ggplot(searchengine, aes(fill = countries, y = values, x = brands)) +
  geom_bar(position = position_dodge(width = 1), stat = "identity", width = 1, 
           color = "black") +
  facet_wrap(~ countries, ncol = 3, scales = "free") +
  theme(legend.position = "none", plot.title = element_text(hjust = 0.5)) +
  geom_text(aes(label = values), position = position_dodge(width = 1), 
            vjust = -0.25, color = "black") +
  ggtitle("Top 5 Most used Search Engines amongst the 15 Countries with the highest GDP as of February 2020 ordered Alphabetically") +
  xlab("Search Engines") +
  ylab("Percentage of search engine usage as of February 2020 (%)")

Data Reference

Reconstruction

The following faceted bar chart addresses the issues from the original visualisation and fixes it here.