How to use Tableau to visualise data effectively

What is Tableau and how does it differ from other popular tools like Excel, Power Bi or R?

Tableau is primarily a data visualisation tool. It has some ability to manipulate and organise data with derivative tools (e.g., Tableau Prep), but its main function is to take data and visualise it to tell a story.

Excel, Tableau / PowerBI, and R are powerful tools, but when and where you use them depends on the task at hand.

Excel can do a lot of what Tableau does, but it’s a generic tool that works well for a quick and simple task. I started my career working almost entirely with Excel and its associated Power Tools. I did all my data management and visualisation with it, but Excel has limitations that Tableau helps alleviate.

Tableau and Power BI are both purpose built data visualisation tools designed to make aesthetically pleasing, interactive dashboards. Not only is it easier and quicker to create visuals in Tableau, but the visuals themselves are far better.

The main reason I like Power BI and Tableau, though, is that they both enable interactive dashboards that can’t be broken by an end user. Since the dashboards are published online (another benefit in accessibility), they don’t show any formulas or underlying data connections. That means that someone viewing or engaging with it can’t “break” it, which happened all too much with Excel.

There’s not much difference between Power BI and Tableau. They both have strengths and weaknesses that the other might be more suited for, but it’s very much like PlayStation and Xbox. They both do the same thing and which one you choose doesn’t really matter – you will likely just choose one and stick with it.

RStudio is a slightly different tool in that it can do everything, and do it well, but because it’s a coding language it can be quite daunting and difficult to learn if you’re using it for all the above functions. I use RStudio a lot for data manipulation and some data visualisation, but creating and hosting interactive dashboards through Shiny (an RStudio package that powers interactive visuals similar to Tableau) is just not something I have learned well enough yet.

How does Tableau’s learning curve compare to other data analysis and visualisation tools? Is it more intuitive than other tools that staff members interact with?

I found the learning curve was initially quite steep. Before learning Tableau, I was used to Microsoft Excel and its associated tools liked Power Pivot and Power Query. I found it hard to transition initially as Tableau works in a very different way. It took me a few months until I really felt comfortable.

The main two areas I struggled with initially were loading in multiple databases and getting the relationships between them working correctly; and learning how level of detail (LOD) expressions work. The former just took a lot of trial and error with data structures until I found one that worked well for me. LOD expressions were hard at first simply because you’re learning a new syntax. They are similar to learning how formulas work in Excel. Once you get a handle on a few, it makes more sense.

This shouldn’t dishearten anyone from trying to use Tableau. You can make lots of progress early on by keeping things simple. The time spent learning the system will be worth it in the end.

Do you have any principles that you live by when creating visualisations in Tableau that allow you to communicate data effectively?

Everything has to be made with a purpose. I’m not a big fan of leaving things open ended and up to interpretation. It’s too easy for the reader to get lost or interpret something incorrectly. I try to tell a story with the data that’s easy to follow.

Moreover, designing a report that has no message is difficult since you have nothing to narrow your focus. Choosing a particular metric or the specific colours in the visualization might be hard when you don’t know what you want the reader to notice or take away.

Figure out what the message is before creating the visual.

The most frequent roadblock when creating a visual is deciding what’s important and what’s not. It can be hard to narrow down the focus of some visuals, because many times you want to include all information possible. But creating visuals that cover all the bases is very hard. More importantly, it’s hard for the reader to use and understand.

Keep it as simple as possible and make the main thing the main thing.

Next, context is king. It is easy to misinterpret data when you don’t have all the facts. This is one of the reasons why keeping it simple is hard to do, since adding context means adding more data or visuals.

I always try to use data that is grounded in something. That typically means showing how some data differs from previous training or match norms. Adding that level of context helps the end user grasp whatever message you want to portray.

Our matches are typically the hardest part of our week, so if someone is exceeding their match norm in a particular metric then you know this could be considered “hard.” The alternative is showing the absolute numbers without context – such as 5000 m of total distance – but the viewer might not understand what that actually means.

Last, look up examples of data visualisations online for inspiration. I often go to the ggplot gallery for R studio or Tableau Public to get ideas on how I could communicate the data I have. Lots of people have already solved the same problems or challenges that I face, so make use of these resources.

In pro sport we are all dealing with an increasing amount of data from GPS, force plates, subjective data, injury data etc. How does Tableau assist in clarifying what to look at?

We collect a lot of data within sport and it’s not often that it is all available in one place. We sometimes have data that is collected through a specific piece of equipment and then only explored or exported through that company’s website; or it could be collected by an individual within a department and they become the gatekeeper for that data.

Athlete management systems (AMS) try to solve this, but their visualisation capabilities are often restrictive. However, if you can create a shared database for all the relevant sources (through a cloud based server or otherwise), Tableau can easily become a one stop shop for all visualised data. This puts all the data and the visualization in one place, so you can view them together and in context.

For example, GPS data and availability data can be visualised and aligned together so that context on a player’s modification / injury status can be provided alongside their training load data. Or body composition data can be presented alongside countermovement jump performance.

It might take a couple of weeks to get all of the data sets you’re likely to use together and talking to each other. But the end result is that you’ve essentially built your own bespoke AMS, and that’s a powerful tool inside performance departments.

An additional benefit of Tableau is that site administrators can limit the visuals that some people see. A coach might not need to see all of the visuals relevant to the medical and athletic performance practitioners, and vice versa. By limiting what each individual sees you’re less likely to have information overload.

Could you give a couple of examples of dashboard or reports that you have created in Tableau and a brief story behind them?

Speed dashboard

We wanted to easily see how long it has been since individuals have hit certain speed thresholds, and also an individual’s maximal velocity per week over time. There’s  a drop down option to see either % Max velocity, Max Velocity in m*s^-1, or number of 90% entries.

Rugby player profile

I was challenged with showing a snapshot of a player’s performance in a particular season. The radar plot is not easy to make in Tableau, as it’s not a standard graph option.  But the impact it has as a visual was worth the effort to make it.

Gym lifts

I wanted to create a very simple way for athletic performance coaches to view a player’s lifts over time. This dashboard makes it easy to flick between players and choose the particular exercises – up to three at a time, each presented longitudinally – that you want to see for that individual.

Would you recommend a practitioner to learn Tableau, and why? Do you have any recommended resources for learning Tableau?

That depends on the time you have available. If you do have time and you regularly work with data visualisation, then, yes, I would definitely recommend learning Tableau. It can save you a lot of time in the long run and the visualisations are professional and can be very powerful. Even just having the ability to interact with reports and dashboards adds massive value over PDFs.

If you’re already spending a lot of time working in Excel to build dashboards and reports, then learning Tableau could definitely take you to that next level of reporting. This doesn’t mean that you’ll never touch Excel again – I still use it regularly for other smaller projects. But in the long term, Tableau (or Power BI) is a good tool for practitioners to become experts in.

On the other hand, if you don’t have a lot of spare time to learn how to use Tableau, then you will probably get frustrated with it early on and are less likely to see the benefits. Excel is still a powerful tool and since most people already know how to use it, it saves you the time of learning a new skill.