Data visualisation: It's not only about being good, it needs to look good!

Are you really paying attention to your charts, maps and all other of sort of images? Our research outcomes might be mind-blowing, but people won’t give them the expected value if your graphics are poor. I found myself eager for data visualisation since the start of my career. It is a true mix of art and engineering. Colour palette, line thickness, font size, chart type, all need to match to make your image worth a 1000 words perfectly.

Nowadays, we have a plethora of software and libraries to help us with data visualisation. Excel does an excellent job when your data is not too complex, and you need just a few figures. But to create a map or change a single property in tens of charts, plotting libraries like matplotlib do a better job. Need to publish on a website, Highcharts and Google Charts APIs are very good and allow a lot of interaction with the figure.

Chart’s legend and caption are also crucial information you want the reader to see. Sometimes, it is worth making your axis limits longer to give more space for the legend in a corner. Axis ticks and gridlines help to extract information more precisely. Check if your map is colourblind safe and avoid bright and saturated colours whenever possible. If you have too many lines but you are only interested in the general behaviour, shades of the same colour work better. Be creative, mix chart types, paint the points according to a colour bar to add a third dimension in a 2D plot, cheat the program to make it plot what you want.

You’ll see. People get very impressed by beautiful charts.