If I had a penny for every time I heard the phrase 'Rules are meant to be broken,' I'd be rich. However, this cliché has no place in BI dashboard design. Having the right dashboard charts and graphs are essential to making your dashboard a useful resource. Your users may not be able to articulate it – since they may not know better – but your dashboard may be the victim of what we call a "chart foul." Chart fouls can be detrimental to the usefulness of your dashboard because they indicate that data isn't being presented the right way, or they mean that your chart type or data labels are confusing your users.
We've been building BI dashboards for the last 18 years for large and small companies. You can probably imagine the number of chart fouls we've come across. Today we've pooled our knowledge and present to you some pointers on major chart fouls to avoid:
1. Choosing the wrong chart type.
Have you ever seen a bar chart with more than 15 data sets? That's a considerable amount of information to display horizontally, so chances are that the graph appeared very cluttered and was difficult to read. You're better off displaying such information as a simple line graph. How about a radar graph showing the qualitative scores of potential hires? It's just as confusing as it sounds (see below), and is probably not a good idea either. So in order to show ranking among candidates, a good old-fashioned table of numbers will do. Choosing the right graph or chart for your data is an art. You should select a graph that is easy to understand and is an appropriate representation of your data. In a previous post on Dashboard Charts & Graphs, we reviewed how to choose the best chart type for your BI dashboard.
2. Improper use of data labels.
Data labels should not be a distraction to your graph – or worse – cover any area of your graph. For instance, suppose you would like to display the average monthly temperature for three cities on a line graph. Your best bet may be to use a combination chart which includes your traditional line graph, backed by a table with all the temperature data. Remember that data labels should provide supporting detail and allow you to draw a conclusion more quickly, not add clutter or confusion to your graph.
The left graph features improper use of data labels, cluttering the graph. The right graph leaves the data labels off, allowing the user to properly view the trend lines, and leaves the data in a table underneath. Sometimes "more is more."
3. Your chart does not tell a story.
Dashboard charts should be more than pretty pictures. Sure you can animate your charts – maybe you want a speedometer that sweeps, a bar chart that grows, and a pie chart that fans out until all the values are displayed. Your charts need to provide useful nuggets of information. They should tell a story, show trends, highlight relationships, and ultimately provide actionable information for users. Effects and animations are the icing on the cake. Ensure that your graphs and charts have substance first; you can always add the fluff later.
With these pointers in mind, you’re well on your way to creating a widely used dashboard for your organization.
For more information on creating exceptional BI dashboards, take a look at these additional resources: