Visualizing data using graphs and charts

With analytics becoming increasingly important for web products, reporting data using charts has seen a surge on the web. This has brought in the trend of infographics and new Javascript graphic libraries. Many people though get too occupied in visual design and overlook what charts actually fit their case best. What’s really bad is when infographics end up displaying the wrong information.

What was the last time you saw a chart whose content stuck in your memory? Was it due to the fancy graphic/animation? Does your favourite website use elaborate graphic animations?

I am of the opinion that a simple 2D Bar or Line chart is almost always better at conveying a message than a 3D graphic. Software charting libraries/plugins ought to be used judiciously and not plastered all over a report. The rotating dials and gradients on several corporate website homepages are usually just pointless and irritating. The graphics on Google Trends on the other hand are so useful cause they are crisp at presenting data. 10 years back the world wide web used to be full of hotch-potch gifs, flash and blinking orange/green lights on jet-black backgrounds! You got to be cautious not to do that with your graphics.

Here’re my tips on how to go about creating a graph or a chart:

  1. Start with the problem — NOT by picking up a cool graphics library.
  2. Ask yourself if you have enough data. Use good statistical methods to analyse that data. If you aren’t ready to crunch the numbers, find help but don’t do it wrongly.
  3. The decision on which chart to use comes at the end. The choice should mostly be a simple 2D representation with optimum contrast and colours that are easy on the eyes.
  4. Think about Accessibility: Use colours that are readable by colour blind people.
  5. In chasing simple and minimal don’t leave it upon the user to figure out what is being said. That’s cognitive overhead which needs to be avoided.
  6. Here’s an interesting infographic showing the basic types of charts and graphs





Leave a Reply