Data-ink vs. understanding

The concept of the data-ink ratio emphasizes the importance of reducing visual clutter and maximizing the use of ink to display relevant data. Still, it is also essential to balance this with clear and concise language to ensure the message is accessible to all.

It is tempting to increase the complexity of a particular data visualization by adding more information, but the understanding of the data rarely improves with it. At the same time, distributing the data across multiple charts does little to improve understanding, as it is rare for anyone to remember the details that connect them.

To enrich the visualization and enable answering more questions with a single chart, I have adapted the famous "candlestick chart" used in the financial market to show categorical attributes. A candlestick chart is a valuable tool for enriching the visualization and obtaining more precise and detailed insights into the data of complaints from utilities.

Candlestick is a graphical analysis technique created in the 18th century Japan and still widely used today. The chart is named because the bars form 'candles,' as shown in the example. Unlike the original technique that shows the evolution of a stock over each period, our chart displays six critical pieces of information for each utility company to show the evolution of complaints in a moving window of 12 months. To achieve this, I used one 'candle' for each distributor.

In this way, I was able to show the highest and the lowest value reached in 5 years (whiskers), the number of complaints at the beginning and the end of the year (candle start and end), the increase or decrease in complaints (candle color, red for increase and green for a decrease), and magnitude of the movement (size of the candle).

The result is a comparison of 52 companies, their complaints evolution, positions in a five years period, and the magnitude of variation in 2022, all in a single chart. Even knowing that an explanation must ensure some initial understanding and that the chart is probably not intuitive enough to be understandable at first sight for continuous use within an organization, it is a good example (I think) of a good data-ink ratio proportion. See the result below: