Best Practices for Creating Useful Dashboards

Find the Best Chart for You Metrics

Selecting the best chart for you metrics is not always straightforward, but it is more than just personal preference. Quite frankly, certain charts do not visualize some datasets well, and certain charts can not visualize some datasets at all. Knowing the difference is a very valuable position to be in.  We have copied below a chart type decision tree that will be a great way to help get the right chart type. You can use this flowchart to select your visualizations. Please download this, print it out, and put it on the wall near your desk.  It has been and will be useful for many.

We should start with the answering the main question.  What type of data set do you want to show?

There are essentially four types of data visualizations you can show:

  • Relationship: how multiple independent variables relate to each other.
  • Comparison: how two or more sets of data compare with each other.
  • Distribution: how different sets of data are spread over a population or other distribution.
  • Composition: how a set of data is made up by smaller divisions.

Returning to our example, we are currently at a stage in the process that has a number of tables, some of them are alike and some of them are not.  Each of them can be turned into a chart type using the decision tree below. Here are a few examples:

Although “the main” question is still to come, There is possibly one data visualization that you will need that doesn’t fall into one of the above groups.  In some data visualization tools, a single value chart is available. Single values not being one of the above visualizations, the very first question you need to ask is “Do you want to display a single value or multiple values?”  

If you want to display a single value in BI Platforms like Chartio you can show that in one of three ways.  The Single Value Chart, which is just one number. We will be using that chart in this lesson. The Single Value Indicator Chart which, technically, shows the relationship between two numbers but only presents a single number.  Finally, the Bullet Chart, which shows a single number and how it compares to a goal you can preset within the chart settings.

For multiple values, it is best to follow the flowchart to make your decision.


Using this Flow Chart we can go through the decision making process for the two of the tables we created in the previous lesson.

For the Revenue Chart (Metric 1) we only have one number so the decision is made for us in the very first question, we only want to display a single number so we will use the Single Value Chart.

For the Monthly Total Subscriptions Chart (Metric 7) we have a series of numbers that are arrayed over a period of time.  We want to know how each row in the table, or month, compares with the previous and following months. For that, we want to make a Comparison and since it is using a continuous data set, we want to use the Line Chart.

Conclusion

When selecting a visualization for your charts the above flowcharts and principles can help you make an informed decision.  It is important for your advancement as an analyst and dashboard author to trust these types of resources for creating visualizations.  You will not want to force a use case through to create the visualization you may have wanted to see. That is not to say creativity is not important and useful it is just to say you may want to know if the time and place is right for extemporaneous chart type usage.  Again, know your audience and also, know your data.

Additional resources:

The cheat sheet from Andrew Abela, the provost at The Catholic University of America and the dean of the School of Business and Economics, provides a general overview.