Nobody sets out to make a bad graph. But it happens. Again and again. It happens in places where you would expect people to know better.
Let's look at the following example from Khaleej Times published on Oct 11 in the business section:
Storytelling is an art. The art is further exemplified when the storytelling involves data. The story in the article could have been told differently to resonate better with the readers. However, I'll focus on the following chart from the article to discuss the basics of data visualization.
Let's analyze this chart:
The focus of this article was to highlight the inconsistencies in the chart and not analyze the effectiveness of storytelling, which is another interesting perspective to analyze. The article talks about 'Why Dubai matters to the rick' but does not indicate how Dubai is placed in the ranking of world's wealthiest cities. The credibility of the story is lost if it fails to gives the complete picture and shares information selectively.
We have the best tools available today for preparing 'sexy' looking charts. However, storytelling is an art and cannot be automated by the tool. Stories resonate and stick with us in ways that data cannot.
Let's look at the following example from Khaleej Times published on Oct 11 in the business section:
Khaleej Times (Oct 11, 2019 Page 14) |
Khaleej Times (Oct 11, 2019, Page 16) |
Let's analyze this chart:
- This is a horizontal bar chart, which flips the vertical bar chart. The horizontal bar chart is easier to read especially when the category names are long (e.g. San Francisco Bay Area).
- The chart is easy to read because of the way we typically process information - starting at the top left and making z's with our eyes across the screen or page - the structure of the horizontal bar chart is such that our eyes hit the category names before the actual data.
- The chart uses the logical ordering of categories by ordering the cities in descending order (and not alphabetically)
The 'bad elements' that wreck this chart:
- The data is presented in Trillions but the color coding distinguishes it based on Trillions and Billions. The size of Singapore is 1 Trillion whereas Chicago is 980 Billion. It is not very obvious when you read the chart - a higher number (980) is placed below a much smaller number (1)
- The data for billions seems to be entered manually and there is an inconsistency in the data and bar size for Seoul, Paris and Shenzhen (980 billion each) whereas bar is the same as Osaka (790 billion)
- Distinguishing trillions and billions add nothing to the story that the data is trying to communicate. In fact it confuses the reader. Color is a very important affordance in design and makes it obvious to the reader on how the product (in this case the data or the chart) is to be used. Pre-attentive attributes are used to draw the reader's attention to where we want them to focus - in other words to high light the important stuff.
- The title of the chart does not indicate what the data indicates. $ 3 trillion or $ 980 billion means what ? Is it the overall GDP of the city or the overall wealth of the individuals ? It is not very clear from the chart.
- The color used in the chart places more emphases on the countries with blue bars. The hue of peach color has less pre-attentive attribute than hue for blue color.
- The number orientation is not very consistent. Some numbers are mentioned with no decimal (3) and some are presented with one decimal. Changing the number from trillion to billion is a disaster which we have already discussed.
- The message in second part of the article - The Big Apple Rules: New York has 65 billionaires and over 380,000 millionaires as residents is not coming out clearly in the chart. In fact if the person does not read the second part of the article, the message is completely missed.
Let's try and see how we can improve the affordances of the chart:
- The chart title is altered to indicate what the data means
- Sub-heading is added to give context of the data that needs to be highlighted
- Color is used as a pre-attentive attribute to highlight New York city
- Top 10 countries are colored in Blue (other than New York) and the next 10 are de-emphasized by using a grey color
- The data is presented in consistent terms - 2 decimal and one measure of currency is used
- Legend for currency is placed close to the first bar for ease of readibility
- No color legend is given, the color used (orange, blue and grey) signals the emphasis of data
- The color used are color blind safe
The focus of this article was to highlight the inconsistencies in the chart and not analyze the effectiveness of storytelling, which is another interesting perspective to analyze. The article talks about 'Why Dubai matters to the rick' but does not indicate how Dubai is placed in the ranking of world's wealthiest cities. The credibility of the story is lost if it fails to gives the complete picture and shares information selectively.
We have the best tools available today for preparing 'sexy' looking charts. However, storytelling is an art and cannot be automated by the tool. Stories resonate and stick with us in ways that data cannot.