I recently published Beautiful Data Visualization – Part 1 in which I discussed some of the challenges with data when it comes to creating compelling visualizations. I also offered some suggestions of tools and processes to build confidence in your data sets. Now that you’ve taken the first step to great data visualization (or at least formulated a plan to do so) it’s time to start developing those visuals which will provide executives, managers, directors, and the like with the insights needed to make smart business decisions.
The Starting Line
To begin you must have an objective. In many cases these directly answer very specific questions. “How much does it cost us when deliveries are late?” “What is our spend compared to budget for the current year?” Now you can begin to review the available data which will be used for analysis. You’ll be able to identify any gaps that an auxiliary data set can fill and become familiar with data you will be working with while developing. Finally, the more exciting part! Prototyping. You get to brainstorm and design concepts for visually the data. Some questions you’ll want to answer are: What data will you use? What type of chart or visualization is best for your objective? Are additional indicators or comparisons required? Once you’ve set the requirements or parameters you are ready to create a mockup for approval.
As you build your visualization give consideration to making it as natural as possible to be actionable. Many business intelligence tools and visualizations do not provide actionable information. This is what I call linear business intelligence. What we want to create is a natural method of discovery that leads to action. While some of that does rest on the data underneath your visualizations, the latter cannot be underestimated or taken for granted.
This is why it’s very important to consider what type of visualization is best. There are some 21 different chart types that I would consider using. But they each have their intended purpose. Some are for comparisons, others for composition, some for distribution, and a few for relationship (we’ll discuss this in a future post). This is why we start with knowing the objective, asking questions, reviewing the data, and then creating mockups. This is what will lead to the use of the most compelling visualization.
Here is a real life example of a travel dashboard I was tasked to create. I knew that I needed to create a consolidated view of travel spend. I knew that I would need to bring together multiple sources of data. And I knew that I needed it to be actionable. So I started it with three important metrics: 1) Spend vs. Budget, 2) Online booking performance (goal vs. actual), 3) Traveler gamification based on booking habits. The other metrics complement those and lead to further analysis as necessary. The final product was impressive and eventually lead to an entirely new product offering.
- Know your primary objective.
- Ask questions to have a complete understanding of the project and what it will be used to accomplish.
- Review available data that will be used for analysis and to power your visualizations.
- Prototype. Create mockups and get approval from stakeholders before development.
- Post development provide a sneak peak internally or with the project owners and get feedback.
- Adjust as required based on previous step.