Visualization anti-patterns

In addition to report planning and generally aligning reports with visualization best practices, it can be helpful to acknowledge and avoid several common visualization anti-patterns. For many reports, particularly when report development time and Power BI experience is limited, simply avoiding these anti-patterns coupled with adequate planning and appropriate visual type choices is sufficient to deliver quality, sustainable content.

Six of the most common visualization anti-patterns include the following:

• A cluttered interface of many visuals and report elements that's complex or difficult to interpret:
  • This is often the result of too many visuals per report page or too high a precision being displayed
  • Separate reports, report pages, and the removal of unnecessary details and precision can improve usability
• A lack of structure, order, and consistency:
  • Each report page should naturally guide the user from the essential top-left visuals to the supporting visuals
  • A failure to align visuals or to provide proper spacing and borders can make reports appear disorganized
  • Mixing widely disparate grains of detail on the same report page can be disorienting to users
• High density and/or high detail visualizations, such as large table visuals or thousands of points on a scatter chart or map:
  • The need for a scrollbar is a strong indication that a visual contains too many values
  • A table visual should not be used as a raw data extract of many columns and rows
  • High density visuals, such as line and scatter charts with thousands of data points, can cause poor performance

The following table visual with six dimension columns and three measures is an example of a data extract anti-pattern:

Data extract anti-pattern

The small scrollbar on the right indicates that many rows are not displayed. Additionally, the export data option prompts the warning message (data exceeds the limit) suggesting the visual contains too much data.

• The excessive use of fancy or complex visuals and images:
  • Reports can be aesthetic and engaging but the priority should be to inform users, not to impress them.
  • For example, a column chart or a stacked column chart will usually be more effective than a treemap.
• Suboptimal visual choices such as pie charts, donut charts, and gauges:
  • Column or bar charts are easier to interpret than the circular shapes of pie and donut charts.
  • KPI visuals provide more context than gauge visuals including the trend of the indicator value.
• The misuse of colors, such as utilizing more than five colors and overwhelming users with highly saturated colors:
  • Colors should be used selectively and only when the few alternative colors convey meaning.