6 Design Tips for Better Dashboards

In his article "Choosing the Right Chart for your Data," Brian Petersen (VP of Professional Services at Jet Reports) writes: "Data is the foundation of effective business.  ... Being able to quickly read and analyze your data enables you ... to understand how a particular set or group of facts contributes to your overall success and steer your decisions proactively."
He then goes on to describe several common charts including:

• Area Charts
• Bar Charts
• Scatter Charts and
• Stacked Column Charts

Read the full article here:

For more seasoned analysts, the challenge is less about finding the right chart so much as laying out a dashboard that effectively communicates a broad scale of both summarized as well as detailed information. As we delve into optimizing dashboards, we move away from technical or domain expertise and transition toward questions of User Experience and User Interface.

For these projects, I'll leverage knowledge gleaned from Stephen Few's guide to dashboarding -- "Information Dashboard Design" which was heavily influenced by Edward Tufte's seminal work "The Visual Display of Quantitative Information".

Charts Jump off the Page with these 6 Tips

1) "Brevity is the soul of wit" -- Do not exceed 1 page.
Any CxO will describe the perfect dashboard as an interactive report where they can see all the important information on one page. Translation: Edit. Edit. Then edit again. Examine how much excessive detail or decoration you can pare away without supplying inadequate context.

​Your final product shouldn't require scrolling, changing tabs, or (ideally) a legend.

2) How good is good? -- Provide enrichment and context through comparison.

• Same measure at a point in the past (Last Year or Last Month)
• Target (Budget or Forecast)
• Relationship to future target or prior prediction (% of annual budget or forecasted budget)
• Benchmark for norm (company average or industry standard)
• Separate but related measure

3) Consider providing non-quantitative data
If you frame dashboards around improving a process or keeping 'two fingers on the pulse of the company'; in addition to measuring activity, it may make sense to provide non-numeric data.

• Top & bottom performers
• Prioritized issues to investigate
• Upcoming due dates.

4) Emphasize the important things.
By understand how the eye travels across the page, designers can highlight, prioritize and de-epmphasize.  This is particularly important when planning the placement of auxiliary elements including filters, slicers, legends and labels.

5) Maximize the "Data-Ink Ratio"

• If 'data ink' is any line or pixel that communicates quantitative data (a line, point or bar in a chart) and 'total ink' includes all the lines used to create said chart (axes, tick marks, borders), then the data-ink ratio measures the distribution of ink used for communication versus formatting and decoration.
• Strive to enhance data ink while reducing and deemphasizing non- data ink.
  1. Can you remove unnecessary tick marks or grid lines?
  2. Do variations in color provide additional meaning, or is the message equally clear with just one or two muted colors?
  3. Do you really need that 3-D effect or color gradients?

6) Organize information to support interpretation and application

• Organize groups according to departments, entities or use
• Co-locate and subtly delineate related data
• Support meaningful comparison while minimizing meaningless comparisons

My favorite feature of Few's book was his analysis of sample dashboards, wherein he described, not only the flaws in various dashboards but also modeled various alternative ways of presenting the data.  For dashboard developers this analysis would prove invaluable for sharpening our critical eye, but also provide inspiration for what dashboards can (or shouldn't) look like!

Bridge the Gap between Concept and Execution

For those of you using Pivot Tables or PowerBI to access data from a Jet Enterprise cube, it can be difficult pursue the optimum dashboard layout or chart because you're constrained by the limits of the pivot table or data model.
In a previous post: Better Dashboarding with Cube Functions, Onyx Reporting presents a tutorial for converting Pivot Tables into infinitely more manipulable Cube functions.

One Book to Rule them All

Content and images from this blog post were taken from Stephen Few's Information Dashboard Design (buy it on Amazon).

Check out our blog at www.onyxreporting.com/blog or sign up for our weekly newsletter: "The Data Dump".

If you need support knocking out a batch of reports or want to customize your Jet Reports cubes to include some new comparative measures, our services team is amoung the best in the business.