How to design a good Tableau Dashboard – quick guidelines and best practices.

tableau online live dashboards

When answering a question, you must consider that along with the actual answer, it is important how you answer the question, thus when creating visualizations to answer business requirements you must always keep in mind to present the information as if giving a nicely wrapped gift, the gift of data visualization using Tableau Software.
So, how do we wrap the gift the of data visualization nicely? Here a below a few guidelines and best practices we have encountered and applied over the years:

1. Start with the Tableau Dashboard
When answering a business requirement always start the design from the dashboard, not from the sheets or data. It is easier to design a layout, query the data, choose the charts, select colors and fonts and review the dashboard than connecting to data, creating charts and trying to find fit the charts in a dashboard.

2. Dashboard design guidelines
When designing a dashboard, we have found it useful to stick to a well-defined set of rules:
– Structure the dashboard so that the most important information to display is in the top-left corner and organize the chained interactivity either from top to bottom or from left to right.
– Always try to keep the number of visualizations in dashboard to a maximum of 5-6. You can always navigate to another Tableau dashboard to display further information if needed.
– Avoid using multiple color schemas in the dashboard.
– Group the filters together and emphasize them with a light border or shading. The left, right or top is usually a great place to put them. Also, any legends that are generally applicable can be grouped together with filters.
– Avoid scrollbars in the main views of your dashboards.
– Size your dashboard and views based on the device that will be used to view it (mobile, tablet, laptop, desktop-PC).

3. Choose the right chart type for your visualization
We continue with the charts, when showing a certain type of analysis choose the best visualization. For example, trends over time should be represented by a line, area or bar chart, comparisons should be represented by bar charts, correlations should be represented by scatter plots, distributions should be represented by box-plots or histograms, part-to-whole should be represented by stacked bar charts, etc.
Sometimes, from our experience, you will combine multiple chart types in a dashboard to create one visualization, but we will not dig that deep for now.

4. Create effective views using Tableau Software
Even after choosing the best chart type it’s not all done. The first thing to consider is to emphasize the most important data from the visualization, through color, shape, size or label. Next you should make a view easy to read, focus on orientation of both labels and charts. Furthermore, if there are multiple dimensions in the view consider positioning them so that the visualization looks clean and neat.
Once these steps have been performed you must check for the following: avoid overloading your dashboard with elements (reference lines, shapes, colors, size) and limit the number of colors and shapes in a view.

5. Perfecting and evaluating your creation
Once the dashboard has been created, it is healthy and necessary to review it. Here are a few areas to always at:
– Color: look to see if colors are obvious and self-explanatory, check to see if there are overlapping color scales, do not use distinct encoding of colors for more than 10 values.
– Fonts: select a font that is readable, consider using a different font for tool tips as you can emphasize them this way, investigate color coding fonts to display more information through color of fonts.
– Tool tips: they should bring more information to the user and if should be self-explanatory as there is no tool tip for tool tips.
– Axis: ensure you use axis properly, if the information displayed is not relevant and you may end up saving space consider removing unwanted axis.
– Labels: if there are too many points in your visualization, consider showing labels when a user selects a mark or when a user highlights a dimension or show min-max, first-latest values.
As a final note, the aim of any design should be to create something beautiful, inspiring, interactive, holistic and in our case intriguing, so that users will feel engaged with the Tableau Software dashboard and think of new questions based on our answers.

Download Tableau Software