In numbers we trust?
Our business life and much of our experiences are grounded on the credence to numbers. We trust numbers as the means to reaching truth and objectivity in analysis, as well as confidence in conclusions. They tell us about the variety of important subjects, moreover, they present a distinction between what is accepted as harmless and what is supposed to be dangerous. And since data-driven decisions define the further course of actions we with the same rigour long for objectivity through numbers in business reports and in health issues.
Throughout recent blogs, we’ve discussed reasons why data visualization is an excellent tool to convey complicated information in an appealing way. By visualizing information, our brains can integrate and maintain complex content more productively, enhancing its influence. But if data isn’t accurately reflected, it can do more harm than good. The faulty presentation can wane the data’s report or, worse, distort it completely.
Here are 5 popular data visualization mistakes which lead to misinterpretation and distortion of the information presented throughout different data visualisation tools and techniques.
Choosing the inaccurate visualization
First of all, by choosing a visualization based on your personal taste rather than on the nature of your data you might overload your presentation and make it harder to interpret. There are a lot of situations when bar chart can do a much better job than bubble chart and scatter plot can show your message about the relation between variables in a nice and clean way. Also, you might have noticed that sometimes a simple table is the most optimal way to present data than any colorful chart.
Try to keep in mind you goal and needs of your audience, this will help to avoid unnecessary intricacy.
Designing Data Non-Intuitively
A mere “5-second test” may help you to see the flaws in your data visualisation. Can one easily grasp the main idea in a few seconds?
Such elements as non-solid lines in a line chart, mixed colors on a heat map, misordered pie segments, non-calibrate time series must be avoided due to a bad impact on perception. Do not make the user do more work. By trying hard to be original you may make it arduous to see and investigate data.
Remember a simple rule – always make your purpose as clear as possible!
Omitting and misrepresenting data
In the pursuit of the insights and analytical findings, it is possible to get into a mess with omitting and misrepresenting data.
Such elements as truncated Y-axis, omitted data, unclear legend, an over-dramatized small difference should be dangerous flags for you.
For instance, with the help of such tricks, one can mask the true volatility of the market in the business report to make company’s profits look steady and predictable. But even without such intention one can easily fail calculations. That is why it is crucial to have deep knowledge about your data.
So, never compare carrots to bananas, don’t alarm your CEO unnecessarily and pay attention to the validity of the representation of your data.
Presenting too many data
Even with all available dashboards tools, one still can try to force any possible information into one chart. More is better, is not it? Not always.
Over-complication is a data visualization error that is sure to scare most viewers as it makes it difficult for them to understand where and what to focus on.
Do not forget, data visualization is meant for clarifying and structuring massive amount of data, not vice versa.
Displaying naked data with no degree of certainty
One of the most popular objections of executives is that business analysts throw them in data masses with no paddle. Data visualisation is not enough as long as it does not help recognize the degree of certainty if an action you are advising will produce desirable outcomes. Also, if you ever wondered, there is such a thing as small data. And the results of the analysis with such data can be not a disaster but close.
Therefore, it is always a good time for a course of statistical business analysis. You may not need a deep program but implementing even few basic element in your reporting routine can change it all. At least, you might revise popular but sometimes futile KPI metrics such as average, percentage etc.
I hope listed mistakes had proved to you that sound data visualization is more than just picking the decent chart type. It’s about offering information on a plate as a main dish both beautiful, tasty, made from qualitative ingredients with a proper flatware adjacent.