Over the past few months, one of my responsibilities at the Data School Down Under has been to review Tableau dashboards. This task has raised the question of how I approach building Tableau dashboards. In today’s blog, I wanted to share one of my preferred approaches, and provide some general areas of discussion to take your Tableau dashboards to the next level.


Step 1: Understanding the data

The first thing I like to do when thinking about my approach to Tableau Dashboards is opening Tableau Prep Builder. The great thing about Tableau Prep is the profile pane, which very quickly allows me to see and understand the underlying data. This is especially cool with the inbuilt highlight action, which will allow you to see distributions among different fields.

While this is a quick way to do this, you can also do similar data discovery within Tableau Desktop, dragging in fields onto the canvas and seeing how they interact with one another. Including counts will show distribution.

The next thing which I have found works for me is to bring out a pen and paper, and group all the fields I have by DateTime, Dimension, Measure, Spatial and Other. Not only does writing down all the fields help me remember what I have available, but it also helps me clarify which fields I want to use and how these will relate to one another. This will also give me a good understanding of the charts I may want to use to explore in more detail, which brings me to…


Step 2: Experimenting with the Charts

After my initial understanding of the data (e.g. granularity and data types), I will usually go into Tableau and build the following fundamental charts:

  • Bar Charts (Comparing Dimensions to Measures)
  • Line Charts (Comparing Datetime)
  • Map (Comparing Spatial Objects)
  • Scatterplot (Relationships between 2 measures)
  • Crosstabs (This is more for checking numbers)

Conversely, there are some charts that have political incorrectness, namely charts that don’t use visual best practice. These are pie charts & treemaps (usually better as a form of bar), large crosstabs (can implement a highlight table or another chart), area charts (can be used effectively but hard to compare with multiple stacked categories), and bubble charts (usually a bar is better in this instance too).

If there are no strict requirements for what the Tableau dashboard wants to show, I try to stay active by asking myself what the chart is telling me, why it might be telling me that, and then ask myself further questions I might want to know. I have found at the beginning these analysis questions may not be clear, in fact, it may take a few attempts experimenting with different charts and Tableau dashboards before you find an interesting story.

If however there are strict requirements then the approach would be slightly different. Before experimenting with different charts, I would want to think about what value the dashboard can make, and I would go about drawing dashboard layout(s) to answers these questions, and I would then evaluate whether the dashboard supports the story/goal.


Step 3: Building the Functionality

After experimenting with a few charts and dashboard layouts, I would decide on the layout, build all the interactivity, and create my tooltips. Interactivity would be in the form of Tableau actions. I would then click around the dashboard and at a functionality and interest level decide its value.

Tooltips are useful to provide some extra context into the visualization, and preferably should be written in an English sentence. You can also create a viz in the tooltip, which is demonstrated in a previous blog.

If I do not like it, I would normally duplicate the dashboard and either make minor modifications to capture an actionable dashboard or re-think the layout and the story I would want to explore. This would continue until I am happy with the visible functionality.


Step 4: Formatting the Dashboard

At the Data School, you get exposed to several different dashboard styles, and you get the opportunity to explore your own personal style. While I cannot confirm that I prefer a strict style, I have noted that there are “better” ways to format, which are supported by visual best practice.

You can find a great report that Tableau has published about visual best practice here 

Some of the tips I use are blank objects and containers to support whitespace, I like to turn off gridlines and chart dividers usually to keep the chart noise low, recently I’ve been experimenting with the map options to support the story. Another cool thing you can do is use Mapbox and import that into Tableau.

The final comment I will make to formatting is that it takes practice and probably the most time-consuming component of the Tableau dashboard. This is because paying attention to the finer details can make a big difference to the overall dashboard appeal. For example, showing 2 decimal places on an axis label is not appropriate.


Other Tips & Tricks:

Data School Australia is always open for applications, and I wanted to end this blog with some general comments or functionalities to take your dashboard to the next level:

  1. Double Encoding: Double encoding is the process of using 2 visual ques to represent the same factor. For example, using color and size to represent a measure on a map. This idea is not always labeled as bad or good, but you want to be as clear as possible on a Tableau Dashboard, and you are essentially missing out on an opportunity to show something more. Be careful however not to over-clutter the chart.
  2. Top / Bottom Filters: In the filters shelf, you have access to create the top or bottom numbers of categories which reduces the size of a dashboard. You can also add these as a set. Depending on how the calculations were built, you may need to add this to context, but it is a good way to reduce scroll bars on a chart.
  3. Filters / Legends Location: Tableau automatically places filters and legends on the right-hand side of the screen. This is because it is normally the preferred location, however, I have also seen filters work well under BAN (Big Numbers) / The Title. Another trick you can use is placing them in a floating container, and within the container drop-down arrow, select add/show button. This will make the filters in a hamburger to save the screen space.
  4. Help Icons: Sometimes a dashboard requires some explanation of how it works, yet this information itself is not important. You can add a dummy calculation like “help”, and place this onto the shape marks card. Within this tooltip, you can add some instructions to support the user’s comprehension of the dashboard.


Overall, the more practice and exposure to dashboarding, the more it will differentiate your Tableau dashboarding skills. Tableau Public is great to show what the community has made, and this can prompt ideas and technical skills, but you should always think critically about what others and yourself make.