Decision-makers depend on insightful charts and graphs to help them make fast, accurate decisions. "Insightful" charts and graphs are easy to read and understand and are the right design for your data. The wrong chart type can throw everything off and make your dashboard unusable. For example, while pie charts are most often used to display the share or percentage of a total, they're not good for comparing the relationship between two variables – a scatter chart is better for that. While bar charts are good for comparison, if you want to compare many categories of data over time, go with a line chart.
Are you tasked with creating the dashboards that are used every day at your organization? As dashboard experts, let's take a look at a few practical examples of why certain charts are better suited to display certain types of data versus others. If you find this article useful, be sure to view our webinar recording Scorecard & Dashboard Development: A Detailed How-To, where you'll get more practical tips to create interactive, "go-to" dashboards.
Sales data for the year may be best displayed on a bullet graph. This type of graph displays a fair amount of data in a small space, compares measures to enrich its meaning, and is generally a simple, uncluttered visual representation of data. With a bullet graph, the budgeted sales amount for the year is represented by the entire length of the bar; the actual value is represented by the thin bar in the forefront (blue on the sales bullet graph). The shaded grey areas (in our image) represent the values "poor," "satisfactory," and "good." In our image, the sales data is satisfactory, approaching good. Stacking bullet graphs allows the user to compare values with ease – here, sales vs. costs. A bullet graph would even work well as a desktop widget since it can showcase an important KPI in a simple, space-efficient format.
Would a bar chart work for this same data?
A bullet chart is actually a variation of a bar chart. Even though bar charts are useful for comparing data as well, for our sales example, it may not be your best option. With a bar chart, you’d need to have 2 columns per month – one for actual and one for the target/budgeted amount. Alternatively, you could use a stacked bar chart, but that is basically what we have here with the bullet graph, only with even more data than a stacked bar chart typically offers. Because it would be difficult to track your progress over time, a standard bar chart isn't the most efficient chart type for sales actual vs. budget data.
Candlestick Charts vs. Bubble Charts
Ever wondered why a candlestick chart works so well for stock market prices? It's because this type of chart is highly informative, showing multiple pieces of information about your stock. You can see the opening price, closing price, volume of trading and the date all on one graph. A candlestick chart also gives you a good feel for the changes in stock prices. Judging from the size of the bar, a quick glance at the graph will tell you whether a change in stock price was minimal or significant.
What about a bubble chart?
Bubble charts show a lot of information as well, but here's why they don't quite work for stock prices. Bubble charts are good for getting an overall sense of comparison versus viewing specific qualitative data. Say you were to put time on the x axis, price on the y axis and make volume the size of the bubble. You'd have to scroll over the bubble to get a pop-up of the price and volume sold, since the chart is out of room to show you that data. Price pattern analysis requires specific information, so a bubble chart in this case would not suffice. It would be excellent though to display the volume and margin of sales by product – again, data that needs visual comparison, not specifics at first glance.
Spider Charts vs. Bar Charts
Let's consider one more example. Spider charts (or radar charts) may seem a little scary and confusing at first, but take a closer look at how useful they can be. These charts display three or more quantitative variables in a two dimensional chart. In a recent market study on mobile BI (Feb 2011), Howard Dresner uses a spider chart to effectively compare the platform coverage, feature support, and total score of 20 companies on the same chart (see right).
Why not a bar chart?
The same variables could have been displayed on a stacked bar chart, where the total would be the full length of the bar and the number of features and platforms covered would be shaded within the column; but with 20 companies, it would turn out to be a very wide (and inconvenient) graph. A spider graph in this case is able to accommodate a much larger volume of data in a small space, while the stacked bar chart would be better suited for smaller data sets.
As you can see, your data is most useful when presented in the right graphs or charts. If you'd like to test different chart types with your own data to see what works best, download the arcplan Enterprise Free Trial. For 30 days, you'll get to see what graphs work best with your data and you'll get to format and animate them however you like, then share them with your co-workers.
For additional information on the various chart types available for your BI dashboards, check out these additional resources: