A Visualization should never be just a nice illustration without a message. Rather, it should always communicate a point. Especially when talking about data visualizations: they should always deliver additional business insights. A great data visualization can be done easily with only a few steps.
Data visualization is a hot topic at the moment, and has been so for a while. Like all things up in their hype curve, so has the reasoning on why data visualization is important and what it actually means, been clouded behind the hype-cloud of meaningless chatter.
The raison d’être of data visualization, however, is very simple: it is to make communication more effective. The basics, therefore, on creating an effective data visualization are also very simple, and can be even narrowed down to three easy steps (in reality, of course, things are never that easy).
Clarify Your Question And Know Your Data
There always needs to be a reason to build a visualization. They are not just funny fillings in a presentation, they always need to give an answer to a question. Visualizations can answer much more complicated questions; from a graph we all are able to see correlations, exceptional values and patterns much faster than from a regular data table.
To be able to formalize your question and to be sure to be able to answer it you need to know your data. That is the start. You need to understand what each metric in your data is; what it measures, what the dimensions are, etc. And of course, it’s good to know at least a bit of the business background, which helps you to understand the connections between variables and the end results.
Calculate Key Numbers and Pick Your Charts
In most cases the numbers you want to show aren’t the one you have measured. In this point it’s even more important that you know your data, so you know which metrics you are able to combine and what kind of calculations you are allowed to do. Most often basic percentage calculations are the ones you actually need, and the higher level mathematics have been done way before the visualization process has begun.
When deciding how you’re going to present your data, you have to make a choice between many possible chart types. You can start by testing different kind of charts, and try to see which one is the best. People are used to see for example time series in a line chart, but it’s not bad idea to try something new at times. Picking a right chart is not always an easy choice but try to think how you can make it easiest to read. People are more used to some charts (e.g. bars, histograms, lines) and can find answers from those really fast. But in some cases for example basic maps or tree maps are very useful.
Highlight your findings
Once you are almost done with visualization you need to tune it with colours, labels and layout. You need to drag viewers’ attention to the right place, people don’t have time to search the most important things. With labels, legends and texts you can make viewers’ life as easiest as possible. Especially in case you have found something really interesting it needs to be highlighted. Colours are something you need to be careful with, some colours are associated to be positive/negative.
Practise makes perfect!