When it comes to data visualisation, the majority of the attention is focused on the tools themselves. What chart type, for example, is appropriate for the data? What is the most effective technique to visualise survey results? These are all important questions, but you should focus as much, if not all, of your attention on what has to be done BEFORE you even open Tableau. Here are five of the most important pre-visualization considerations I recommend in regard to data and visualisation, based on my experience at the Data School.
Identify the 5 W’s
One of the most crucial aspects in creating a dashboard is determining what questions it must answer. The question should not be too broad; instead, it should be short and simple. The simplest technique would be to answer the 5 W's: What, why, where, who, and when. These are the most fundamental and critical questions you should answer. Although answering all questions may not be necessary to achieve your goal, doing so will assist you in analysing data from all angles and in answering the core questions. This will also assist you in determining whether adding additional data might improve the quality of the analysis. Remember that every decision you make on your dashboard should be related back to that critical question (or the relevant W's).
Prepare Data using a Data prep tool
Now that you've determined the critical questions, it's time to prepare your data. Preparing your data is crucial since you may need to remove trailing spaces, nulls, change the data type, pivot, and so on. You can do this using a variety of tools, ranging from Alteryx to Tableau's data prep. Whatever tools you use, remember that this is an ESSENTIAL STEP. Many people waste a lot of time working with data that hasn't been prepared. Have you tried to visualize pivot tables? You wind up with extremely long formulas that take far more time to type than is required. A quick pivot before visualising can save you days of effort!
Survey data is one of the most common examples of this. Every response is a row, and each question is a column, with individual responses at each intersection. If you do not restructure your data and pivot the survey questions into rows, you will end up creating ridiculously long formulas to determine average response scores, and you will incur significant long-term maintenance costs when survey questions change.
Keep it simple silly!
Once the data has been correctly structured and the critical questions have been established, it is time to create the charts. It is important to realize while producing charts that it is not required to utilise multiple chart types to answer different questions. The optimum chart type is one that is simple to comprehend. When people are new to Tableau, they often make the mistake of trying to use different types of charts solely for the purpose of variety or to demonstrate their knowledge on different charts. The use of unsuitable charts to visualise data makes it difficult to interpret the data, which negates the objective of the visualisation. Showing the chart to a friend or family member who knows little to nothing about the subject matter or Tableau is a simple way to see if it's easy to understand. If they get it, it indicates you're on the right direction.
Sketch your dashboard
Do not create the dashboard right after you finish designing your charts. Only when a chart is appropriately positioned in the dashboard can it be easily understood. If you put everything in the dashboard and then start organising, you'll lose a lot of time. The proper technique to proceed would be to first sketch out how you intend to present your useful charts. This can be done with a pen and paper or using an online tool like excalidraw. It's not necessary to have a dashboard that perfectly matches your sketch because you might come up with new design ideas when putting it together. Sketching the dashboard, however, will assist you in establishing the "narrative" or "flow," as well as determining how you want the audience to navigate the dashboard.
Test and iterate
Presenting your dashboard to different people is the easiest approach to determine whether the narrative or flow you have in mind is well executed. Get feedback on your dashboard and make the necessary modifications. Reiterate, test, and repeat until you are satisfied with the results. The main thing to remember is that a dashboard is never perfect and can always be improved. The best time to end reiterating is when the user easily understands the visualisation.
