We are living in the world of big data. The cost of capturing and storing data has come down drastically. This has led to an explosion in the amount of data available. However, its important to note that more data does not necessarily mean more information. It may very well mean more confusion and chaos.
Making sense of this enormous amount of data is a critical activity and comprises two steps: analyzing and understanding the data and presenting the understanding in an intuitive and easy to understand fashion. It is the second step where intelligent and well thought of data visualization comes into play.
Although a lot can be said about achieving successful and impactful data visualization, I will start by noting down the seven data visualization sins that must be avoided.
- Preferring Aesthetics over Functionality: Graphs/Charts or any other data visualization elements are important functional components of any report/presentation. Their purpose is to aid in story telling, make it easier and less time-consuming for the user to understand data, and make (sometimes not so apparent) trends and insights “stare in your face”. The purpose certainly is not to decorate the report/presentation. It a sticky wicket when someone starts inserting visuals just for the sake of visual appeal . I have seen numerous reports where most of the graphs neither provide any insights nor they aid in understanding the message. They are just there because someone wanted to embellish the report with some visuals (“jazz’em up with a few graphs”).
- Have them, Use them Mentality: With so many visualization tools available so easily, its takes almost zero effort to graph any data. This has led to a tendency of presenting each and every bit of information visually. Sometimes, it’s just better to rely on tables qualitative comments, etc. Don’t plot just because its easy to do so, or you have the software/tools to do so – do it if and only if it adds any value.
- In-appropriate selection of Charts: Not all graphs are appropriate in all situations. Use the right chart, best suited for your requirements. Using the right graphic is an art as well as science. Give sufficient time to think thru this before you start plotting. This awesome “periodic table of data visualization” is an immensely useful tool in determining which data visualization tool to use in which situation. And, once can always refer to the classic chart selection – though starter by Gene Zelanzy.
- Fetish with 3D: In almost every graph I have seen, 3D does not add any value, in fact it makes the graph difficult to understand and interpret. And in worst cases it can lead to deceptive or wrong data interpretation. I believe that one should always avoid 3D.
- Lack of Message/Story line: The soul of good data visualization lies in a clear and coherent story line. Every graphic should have a message, and everything should add up to the story. And believe me when I say this, this is a damn difficult job. You need to be good with data/numbers, good with the subject/domain, good be the charting software, and be a good (and off course honest!) story-teller too!
- Not doing Sanity Check on Data: Some times people are so involved with the presentation aspect, that they do not focus on data at all. And I am not talking about a very thorough accuracy check, but a basic sanity review. Garbage in garbage out rules always hold true and remember, garbage is still garbage, no matter how beautifully it is presented.
- Not mentioning Disclaimers: Last but definitely not the least important, inserting disclaimers. Any element in the visual that can result in the message being interpreted in a way different than intended, should he highlighted in bold. Something as simple as rounding off numbers, changing axis scale, making the size of various elements dis-proportionate to scale, can lead to incorrect understanding.