How to create an effective BI dashboard

You have identified your dashboard's audience, clarified the questions it will answer, and identified the relivant data. See previous article. What is next?

Choose the correct visualization.

Use bar charts when comparing various categories while keeping time constant—for example, the number of EVs sold by country. To avoid possible misinterpretation, start at zero.

Never use a donut or pie chart because the slices are hard to compare. A bar chart can almost always replace pie charts. Check out this video for more information.

Avoid stacked area line charts as, similar to pie charts, the contribution of each item to the overall total is difficult to interpret when multiple items are decreasing and increasing simultaneously. Instead, a summary line chart combined with small charts depicting each item's contribution to the whole can be used.

Use line charts to show change over time. Use descriptive titles, label your axes, and differentiate your selections by using contrasting colors and thick lines. Ensure they are accessible to the color-blind.

Data considerations.

Group relevant metrics. If the metrics depict a time process, use that to lay out your dashboard and improve its ease of use. For example, one way to describe a sales-specific dashboard could be to include it step by step in a sales funnel.

Only include critical data points. Including too many data points will make your dashboard confusing. Your dashboard should be understandable in 5 seconds or less, and any data included should address the specific question your dashboard is designed to answer.

Provide context. Always include an introduction tab in your dashboard describing its use and a data dictionary tab that your audience can refer to if they have questions about how KPIs are defined. Color your KPIs and add a legend so that your audience is immediately aware of what metrics are overperforming and underperforming.

Choose the proper precision or rounding for your KPIs. Apply them consistently for ease of interpretation. For example, income shouldn't be represented down to the last dollar, and if you also include revenue, be sure that both are shown at the same precision.

Lastly, wireframing with dummy data is recommended. Sharing your dashboard ideas early and often with your stakeholders and others will help you incorporate improvements early in your dashboard design process and deliver a better dashboard faster.

What are your recommendations for creating a great dashboard?