Creating a powerful dashboard isn’t just about pulling a simple grid of data from the database and letting it tell its own story. You have to choose the right visualization, format it appropriately, lay it out intuitively, and highlight the important aspects.
Take a look at this graph, Graph A:
Now this one, Graph B:
This is the exact same data displayed in two ways. Both are good, but how would I decide between bars or lines?
There’s a subtle, but significant difference in how your eye observes these two graphs. Notice that in graph A, you’re automatically drawn to the difference between blue and orange – the difference between books and music. The questions I ask myself when I observe that graph are whether that size difference is growing or whether there’s a particular month where the comparison was large/small. What would stand out would be a month where the size gap between the two bars shrinks significantly.
When I analyze graph B, what strikes me is the trend that both books and music are growing with the exception of a dip in April. I see that both generally grow and shrink simultaneously, and a dip in performance by one of them would jump off the page at me.
So, which is it? It all comes down to the users.
If you expect your users to be looking for cyclical trends of boom and bust over time, then it’s got to be the lines. On the other hand, if the business is geared towards growth in books catching electronics or achieving a balance in sales among the two, then clearly you need the bars.
Bottom line: Graph types drive a specific type of business decision, and choosing the right graph type must always take into consideration the business decision you’re looking to inform.
As a sidenote, here are two very bad answers to the question of which graph to use:
- “I hate bars so make it lines.”
- “I love bars so make it bars.”
Does this sound silly? It should, but I hear this kind of thing said all the time. Maybe not in the context of simple bars and lines, but I am certain I’ve heard the following statements more than once:
- “I hate heat maps.”
- “I love sparklines.”
- “We need some kind of pie chart in the corner.”
Certainly, some graphs are overused. Others can overcomplicate the information and cloud the decision, but every graph type has its wheelhouse. Try your best to think through how the user will make a decision based on your graph when selecting the type.
For more information on graphic design and selection, check out our half day Visualizations Best Practices class. We currently have openings for this class available on December 18, January 29, and February 26.
Latest posts by Pandera Systems (see all)
- How Journey Analytics Transcends Traditional Customer Mapping - September 14, 2018
- Success Story – Staples and Pandera build a single version of the truth - September 6, 2018
- Follow your “HEART” when evaluating your data story - August 24, 2018