Best Practices

Before you start creating a dashboard and visualizations, there is some pre-work that should be completed in advance.

1. Define the purpose of your dashboard. What is it supposed to do/convey? (Eg., educate, influence, communicate, make decisions, take action, etc.)
2. Determine the Key Performance Indictors or Benchmarks that will best accomplish your purpose.
3. Choose the metrics necessary to demonstrate your KPIs or Benchmarks. Which metric will help achieve your goal or solve your problem?
4. Choose the visualization or series of visualizations that can best convey these metrics. A full dashboard, a standalone chart, an infographic?
5. Now you can begin to build your visualization/dashboard.

Commonly cited best practices include:

1. Consider your audience:
   • • Who is going to be using this and why?
     • What do they need? What is the intended function for this dashboard?
     • What is their level of expertise?
     • Do not try to make it one size fits all. It should be targeted to a specific goal. Answer the need from the right perspective.
2. Use the right visualization tool (graph, chart, etc.) for the data and message that you are trying to convey:
   • • Know what type of data you are presenting and the tool that is best able to present this type of data.
     • Know how the visualization will help your audience understand your data.
3. Keep it simple:
   • • 3D is not your friend.
     • Only include critical data.
     • Present high level data, and numbers. Round off numbers and limit the number of numbers (i.e. 32 MIL versus 32,000,000). Tables are the best place for the actual numerical data.
     • An axis should always start at zero.
     • People should not have to struggle to understand your data/math.
4. Avoid visual clutter. Don’t add any elements that don’t need to be there.
5. Don’t try and get everything on the one screen:
   • • Convey the high level information on the first page.
     • Use subtabs and drill-downs as necessary to get into more detail.
     • Use navigational aids to let people know where they are.
     • One concept per visualization; Busy graphics are harder to read and understand.
6. Have a colour scheme; don’t go overboard with too many colours. 
   • • Six or fewer is best.
     • Remember that some people are colour-blind.
7. Colour should only be used to enhance the information you want to convey; beware of distracting the eye from the data.
8. Be consistent: naming conventions, colours, icons, styles …
9. Consider the layout: the most important information should be at the top left for English speaking readers.
10. Dashboards are more meaningful when they provide the answers needed to take action or solve problems, such as strategies, trends, deviations from the norm.  
11. Provide context to the visualization. Present the data in the context of your organization’s goals.
12. You are telling a story with data. Make sure your story is clear and meaningful. Keep asking “What’s the Point?”
13. People should be able to understand your data in one glance, within eight seconds. Someone not connected to your data should be able to view the dashboard from a distance and identify trends or outliers  in general terms.

Duarte, N. (2014) The quick and dirty on data visualization. Harvard Business Review.

• 5 questions to ask yourself when visualizing data
• Includes examples

International Development Research Centre. (n.d.) 10 data visualization tips. Retrieved from the IDRC website.

• Tip sheet highlighting 10 fundamental points, with some examples

LAC Group Media & Archive Services. (n.d.). Data visualization tip for presenting your data [Blog post]. Retrieved from the Media Lac Group website.

• Summarizes advice from difference sources (with links)

Fusion Brew. (2017, December 22). 10 dashboard design errors [and how to avoid them] [Blog post]. Retrieved from the Fusion Charts website.

• Examples of ten common errors with dashboards

Richardson, J. (2009, November 19). Tips for implementers: The basics of good dashboard design. Retrieved from the UMSL website.

• Lists 12 fundamental points for dashboards
• Gartner RAS Core Research Note G00171685

Wolny, T. (n.d.). Build a visual dashboard in 10 steps. [Blog post]. Retrieved from the iSixSigma website.

• Outlines the steps a six sigma company took to design and build a dashboard, from asking the questions, gathering the metrics, creating the design, and delivering the product.

Kaushik, A. (2014, July 15). Digital dashboards: Strategic & tactical: Best practices, tips, examples [Blog post]. Retrieved from Occam’s Razor website.

• Describes why the best dashboards are not just visualizations of lots of data summarizing performance, but put it into context with three sections: Insights, Actions, and Business Impact.

National Network of Libraries of Medicine. Tools and resources: Analysis and visualization. Retrieved from the Network of the National Library of Medicine website.

• Tool kit linking to a variety of resources such as blog posts, data analysis, visualization techniques, how to choose the right chart, and examples.

Yellowfin. (n.d.). Data visualization best practices guide. Retrieved from the YellowFin website.

• Highly visual 17 page guide, one concept per page
• Illustrations of concepts from choice of chart to targeting your message

Abela, A. (2008). How to choose a chart type. In Advanced presentations by design: Creating communication that drives action. San Francisco: Pfeiffer.

• A flowchart to help you decide what type of visualization best fits your intentions

Henry, A. (2012). How to choose the best chart for your data [Blog post]. Retrieved from the Life Hacker website.

• Reviews the steps involved picking the right chart to convey your message
• Does not go into specific design details but looks at it from a higher level perspective

Sharma, H. (n.d.). Best excel chart types for different kinds of data. Retrieved from the Optimize Smart website.

• Free 40-page ebook can be read online or downloaded
• Shows when to best use the charts included in Excel

Hamberg, S. (n.d.). Why you shouldn’t use pie charts [Blog post]. Retrieved from the Funnel website.

• Presents types of graphs and charts, and when to use them

Cherdarchuk, J. (2014, September 26). Salvaging the pie [Blog post]. Retrieved from: DarkhorseAnalytics.com.

• How to make your typical pie chart into a better visualization

Microsoft Power BI. (2019, October 27). Best design practices for reports and visuals. Retrieved from the Microsoft Learn website.

• Looks at the basics of design and layout
• Walks through a clean-up of a bad visualization
• Fairly general but does make specific reference to what can be done / how to do it in Power BI

Tableau. Visual analysis best practices: Sample techniques for making every data visualization useful and beautiful. Retrieved from the Tableau website

• 41-page Position Paper that requires a free download
• Has many explanations and examples, ranging from choosing a chart, fonts, sizes
• Is not Tableau specific