There’s a reason you chose to build dashboard rather than slapping together a handful of Excel bar charts. Finding a better way to present data is the whole idea, so break out of the Excel mindset.
Know how your numbers stack up: pie charts are great when you’re dealing with percentages, a creative X-Y scatter plot will display correlation (and outliers) well, and population data often fits into a standard normal distribution-like curve.
Similar numerical values are suitable for a bar chart, but for widely disparate number sets you might have to find a more creative solution. Space and usability is everything here.
Our favorite approach comes from the teachings of Ed Tufte (watch his interview in the Washington Post here).
Dashboards should make people smarter and should be an excuse to celebrate analytical thinking.
We, of course, believe that ‘effective’ dashboards win over ‘beautiful’. An ineffective dashboard, in our humble opinion, is one that looks more like art than information (see a poor example here). Although appealing to the eye, the graphic has gone way too far and fails to provide immediate value.
How should you evaluate visual assets? Ask yourself if your dashboard gives viewers the ability to pick comparative and correlative analysis without much work.