The Blueprint: How to make a great data visualization

A good visualization is a representation of data that helps the audience or reader clearly see what they otherwise might have been blind to, if they had looked only at the naked source.

So, where to start, and how to plan a great data visualization? Here is the strategy that I follow.

Assess the situation

Step one on your data visualization journey is to identify your business goal. Why? Knowing the goal will help you to stay focused on the outcome, and to ensure that your data visualization serves a purpose.

Typically, your business goal will fall into one or more of the following categories:

- To increase ROI (return on investment),

- To increase revenue,

- To increase profitability,

- To improve productivity,

- Other goals specific to your organization.

So how does visualization help? Data visualization enables you to see the patterns, trends, and facts aligned with your business goal. It enables you to better understand what actions should be prioritized, based on your knowledge about the situation.

The value of the insights uncovered with data visualization is delivered via improvements in your business processes. Therefore, it is important to make sure that your visualization serves a specific purpose – for example:

- Improving analysis to predict and profile ROI and its impact on business,

- Improving understanding of customers and their behavior, especially purchasing behavior,

- Reducing guesswork when making strategic product and revenue decisions,

- Identifying the most profitable, as well as the most money-losing, business activities,

- Identifying the true drivers of financial performance and cost efficiency,

- Identifying customer needs and trends.

Understanding your business goal and the purpose of your visualization will enable you to create better and more thoughtful data representations.

Identify assets

Obviously, you will need data in order to create a visualization.You will need to sift through your data assets, inventorying along the way, to identify data that is relevant, available, and that will help you to meet your goal.

Once you are done with the pre-processing of your dataset – loading, transforming and preparing your data – review the descriptive statistics: the total size of your dataset, the number of unique values, time frame, averages and stand deviations, and so on. Find out if there are any outliers. What facts and numbers are key? Can you find individual examples that are representative of the dataset?

This is the part that a lot of people miss: the more you understand the data source, and the stronger your knowledge about your dataset and the business goals, the greater the potential for a compelling visualization.

The visualization should uncover trends, patterns and outliers. It should evoke an eureka moment – that moment when you really see something for the first time, knowing that it has been right in front of you all along. Without this, a visualization is a mere illustration.

Design a targeted solution

Data exploration and storytelling calls forth a new blend of skills: graphic design and statistics. Having skills in both provides you with the luxury of being able to jump back and forth between datasets and the story behind them. It allows you to go beyond bar charts and explore new types of visualizations – and ones that make more sense.

Because of what visualization is intended for – educating, and changing the perceptions of, your audience – the aspect of graphic design is highly important. Designers spend a lot of time working on the visual side of communication, with a focus on optimum layout, color, typography, and other aspects of design in order to ensure that the final outcome communicates ideas and opinions clearly and persuasively to the audience.

Whether you want to or not, you need to design for your audience: their optimum level of detail, screen size, and color and font preferences. Think of a specific person in your audience and design your visualization for him or her.

When approaching your visualization, it is easy to think that almost every graphic would be better as a bar chart. Resist the temptation to be governed by your visualization library. Rather, be governed by an awareness of human perception, and ensure accuracy for when readers interpret data.

Typically, people respond well to representations of data shifts over time and geographic regions. Include these in your visualization when appropriate.

Research by William Cleveland and Robert McGill on perception and accuracy has revealed that charts where data is positioned along a common scale, is interpreted most accurately. Use scatter plots whenever accurate understanding is of key importance.

Finally, a mark of a good visualization is how fast it can be read and what new facts it allows the audience to immediately pick up.

Test and optimize your visualization

Once you’ve identified the goal and the purpose of your visualization, explored the data and assessed the needs of your audience, you are ready to create the first draft of your visualization.

Make the first draft a brainstorming session, and Include everything you can think of. Then cut it down to one key message that meets the purpose and goal of your visualization. Show it to one or more of your trusted colleagues, and ask for their feedback.

Remember to distinguish ideas facts and suggestions. Suggestions are rooted in opinion and vary according to individuals and situations. Many beginners make the mistake of accepting advice as concrete, and thus lose the purpose of the visualization. You must stick with your data and facts for a worthwhile visualization.

If you have received worthy advice, incorporate it into your work. Sometimes this means redesigning or even completely reworking your visualization. Stay true to your goals and purpose. Improving existing work is hard, but it is usually well worth it.

Sometimes it helps to walk away for a few hours, or even days, in order to review the work with fresh eyes. If you have the luxury of time, I highly recommend doing this.

When about to give your data visualization, go through the following checklist first:

- Check all the titles, headings, and callouts to ensure there are no typos,

- Verify formulas and logic,

- Make sure that all the data is appropriately annotated and explained,

- Review layout, and check for consistency across different screen sizes,

- Make sure that your audience has one key message to take away (i.e., the purpose - what do you want them to do?).

All checked? Now you are ready to publish or deliver your visualization!

Are you planning or creating a visualization? What to know how to make your visualization stand out?

If so, you might benefit from the “Visualization GPS” infographic. It contains all the same ideas as those described in this post, but is a more condensed, to the point version.

Sign up here to download it, and refer to it at any time!

#visualizingdata #datavisualization #visualizationgraphic