The task for this makeover is to design and implement static visualisation for the Merchandise Trade by Region/Market data provided by Department of Statistics, Singapore (DOS). The details can be found here.
The proposed data visualisation is available on Tableau Public through this link.
The original visualisation can be seen below:
Generally, the graph is beautiful but confusing. It has a good and consistent color scheme. The use of trellis display with time-series data conveys the intent of the graph.
Nonetheless, there are plenty of areas to improve with regards to clarity and aesthetics of the graph as discussed below:
The title of the graph is not accurate. In terms of total trade (import + export) for 2019-2020 according to the Merchandise Trade by Region/Market, Indonesia should be included in the ‘Top Six Trading Countries’ while Japan should be excluded. Nevertheless, for total merchandise imports, Japan is ranked 5th while for total merchandise exports, it is ranked 6th.
The date of the graph is not consistent. The title indicates that is for the year 2019-2020. However, the date for individual graphs shows the year 2019-2021 while the date for Japan only shows 2020.
The different y-axis scales for the Export and Import can cause misinterpretation. For example, China seems to have higher Import compared to export as shown in the area chart. However, upon examining the 2 y-axes, Export has a range 0 to 8M while Import has a range of 0 to 6M which means that export is generally higher than import.
The different y-axis scale makes it difficult to compare the trade values across countries. Visually, it seems that Japan has higher trade values compared to others. However, Japan y-axis scales has a range of 0-2M which is significantly lower than Mainland China with range of 0-8M and US with range of 0-6M.
Graphical integrity needs to be improved. The y-axis should be in billion instead of million since the data source indicates “(Thousand Dollars)” so the raw trade values need to be multiplied by 1000.
The graph does not have subtitle which can be used to convey intent and additional information. It also does not indicate the data source.
The area charts hide the export and import information and make it difficult to interpret the graph. Additionally, the area charts render 3 colors instead of 2, probably because of adjusted opacity percentage. The third color is when the export and import overlap which again can cause misunderstanding.
The name of the country is not center aligned. The Import secondary y-axis is also too close to the name of the country heading.
The x-axis title is not consistent with the label. The x-axis indicates ‘Month of Period’ but shows the year 2019, 2020 and 2021. Additionally, there are no tick marks which can represent the months.
The trellis display of each country is not rendered equally and evenly. Japan looks too narrow while Mainland China and Malaysia are a bit wider.
The legend title ‘Measure Names’ is not intuitive. The graph does not include annotations to tell interesting data stories.
The proposed alternative design leverages on the good qualities of the original visualisation. It keeps trellis display with time-series data for each county.
With reference to the critiques previously mentioned, the following suggestions are proposed:
The title of the graph is updated to ‘Singapore’s International Trade with Top Trading Partners, 2019-2020’ to make it more consistent with SingStat infographics. Subtitle is added to explain that Japan ranked 5th for imports and 6th for exports.
The date of observation period on the x-axis is set to Jan 2019 to Dec 2020 to make it coherent with the title. The x-axis title is also removed since the labels obviously indicate trading dates.
To apply the rules for encoding values in graph, line charts are used to display the time-series data. The line charts for Export and Import used the same scale to make it easier to interpret the data. Total Trade line chart is also added to show the total merchandise trade while Net Trade bar chart is included to show the difference of Export and Import.
The same y-axis scale is used across different countries. All countries are placed on a single row instead of multiple rows, so it is easier to compare the trade values.
The y-axis is set to 0-14B to correctly represent the actual trade values.
Subtitle is added to convey additional information. Data source is also included to cite Department of Statistics, Singapore.
Line charts were used instead of area charts to easily illustrate the trade values. Green color is used for Export, red for Import and blue for Total Trade. This color scheme is chosen because it adheres to the practical guides of using color in charts and consistent with the graph from the media release of Enterprise Singapore.
The name of the country is center aligned. The legends were placed on the right side and were not too close to the country heading.
The x-axis labels were updated to be consistent with the graph title. Tick marks were also added to represent the month.
The trellis display of each country has even spacing and are consistent both for the line and bar charts. Additionally, the countries are arranged in ascending order by Total Trade with linear trend line.
The legend title is updated to ‘Trade Types’. Text annotations are added to tell interesting data stories and insights.
The proposed data visualisation can be seen below and available on Tableau Public through this link.
The data is available from Merchandise Trade by Region/Market and can be downloaded by clicking on the link ‘Download all in Excel’ on the same webpage. The file (outputFile.xlsx) consists of 3 sheets - Content, T1 which contains merchandise imports, and T2 which contains merchandise exports. The document includes merchandise trade information for more than one hundred countries and regions starting from Jan 1976 up to the present.
For this makeover, Japan, Hong Kong, Taiwan, United States, Malaysia, and Mainland China are the primary focus and the visualisation is limited to the period of Jan 2019 to Dec 2020.
The proposed data visualisation is relatively simple since it only consists of 2 charts – line charts for the total trade including import and export, and bar chart for the net trade. However, the current data source format is not readily available for analysis. The data is not organized into tidy rows and columns with proper field labels. Hence, additional effort is needed to transform the raw data into a tidy format that can be used to easily visualize the proposed alternative design.
Click ‘Add a clean step’ hyperlink from the figure above. From the Clean 1, switch the display to ‘Show data grid’ to easily view the data.
Click on the ellipsis of ‘Subject: Merchandise Trade’ column. Choose ‘Filter’ -> ‘Selected Values’.
Through data preparation, the raw data has been transformed into a tidy format with only 4 data fields namely Date, Country, Export, and Import which can be easily visualised using Tableau Desktop.
At this point, clarity of the line charts has been established. The succeeding steps will enhance the aesthetics by applying the concepts of data-ink.
Based on the Total Trade linear trend, the trade with Japan is relatively consistent with minimal fluctuations. The total trades with United States and Malaysia are on downward trend. Nevertheless, the total trades with Hong Kong, Taiwan and Mainland China are on upward trend.
Based on the Net Trade bar chart, Hong Kong is the top net exporter while Taiwan is the top net importer. Additionally, the value of exports exceeds imports for Singapore’s merchandise trade with Mainland China. The value of imports exceeds exports for Singapore’s trade with Japan, United States, and Malaysia. These observations were also verified by SingStat infographics.
Based on the line charts and bar graph, there are interesting data stories that maybe correlated to significant events of 2020 COVID-19 pandemic.
Trade with Malaysia significantly dropped on Apr 2020 which coincided with the start of Singapore’s circuit breaker period.
Trade with Mainland China significantly dropped on Feb 2020 around the same period when Chinese government imposed major restrictions.
Trade with US reversed from net importer to net exporter on Apr 2020 around the period after WHO declared COVID-19 as global pandemic.
Mainland China become net exporter from Aug 2020 around the time when China’s economic recovery gained speed.
Note though that these correlations with significant events do NOT imply causation on Singapore’s International Trade. There are many micro- and macro-economic factors that should be taken into consideration to fully understand the dynamics of international trades.
For attribution, please cite this work as
Dolit (2021, May 30). Visual Analytics & Applications: DataViz Makeover 01. Retrieved from https://adolit-vaa.netlify.app/posts/2021-05-30-dataviz-makeover-01/
BibTeX citation
@misc{dolit2021dataviz, author = {Dolit, Archie}, title = {Visual Analytics & Applications: DataViz Makeover 01}, url = {https://adolit-vaa.netlify.app/posts/2021-05-30-dataviz-makeover-01/}, year = {2021} }