Data is better understood when presented in a visual format rather than in text. So how do you choose the visual that best captures what your marketing data is trying to say?
In this post, I will cover the key considerations behind choosing a good visualization.
Your choice of visualization affects the story your data will tell
Data visualizations capture any task that has been measured in the customer journey. You are supposed to organize the observations of a dimension or measure into a graph. But the right visualization option doesn’t always appear immediately when analysts work on their data solutions. Solution menus and dashboards often contain graphics representing the platforms they were meant to measure. These options can work, if you use the tool constantly.
However, analysts often need to combine data in one platform with other data or into a calculated scale. This will change their visualization options. They do not lack options but the growth of data has led to an increase in the number of visualization options for displaying results and integrating data in real time.
All this makes choosing the right visual image more complicated.
Related Article: How Data Visualization Tools Make Self-Service Analytics Easier
To begin defining your perception, ask Munzner’s questions
So where do you start to pick a good graph?
In my Visualization post, What Makes a Good Data Visualization, I mentioned two aspects of data to consider. You want an infographic that conveys ideas from data that is too complex to explain in words and that helps your audience quickly analyze information and act on results.
To access this graph, ask a set of questions created by Tamara Munzner, a professor of computer studies at the University of British Columbia. Munzner is known for her extensive research in the development, evaluation, and characterization of visualization systems and techniques. She highlighted this question framework in her presentation on Avoiding Perception Analysis.
- Who are the end users? (This is the audience that needs the information.)
- What is displayed?
- Why is the user looking at it? (Questions 2 and 3 aim to highlight what the data is, how it is arranged, and where it came from.)
- How is this displayed? (This is the main question – what kind of graph displays the data best.)
Answers to Munzner’s questions help narrow down the graphs that better represent the answers visually. Your choice of graph must fulfill one of the following purposes:
- To analyze the composition of the distribution or change.
- To identify patterns or trends.
- To detect Objective 1 and/or Objective 2 within a subset of a given data set.
Choose a graph that shows a hierarchy in the data
Four classes of graphs are suitable for displaying hierarchies in data: composition, distribution, relationship, and comparison. Both composition and distribution graphs address the structure of dimensions or measures defined through observations, while relationships and comparison graphs aim to highlight contrasting differences through patterns and trends.
Configuration diagrams are intended to describe the composition of a set of notes. Visuals in this category include pie charts, hierarchical maps, and stacked bar charts.
Distribution graphs display the range of observations, making them ideal for statistics indicating the quality of the dimensions and metrics containing those observations. Examples such as histogram or boxplots are chosen to manipulate the statistical range.
Relationship graphs are about trends in correlation between two or more dimensions, or measures. Scatter charts and bubble charts are good examples.
Comparison graphs are intended to highlight differences in terms of deviation, trends, or order between two or more dimensions or measures. These are often a specialized variety of relationship or fitting graphs, such as regression charts, Pareto charts, terrain charts, and stacked bar charts.
The best graph for your purpose organizes the data to answer the question “Why is the user looking at this?”
Each of these categories has multiple graphic styles, more than can be covered in a single post. But when choosing a graph, you are looking for the graph that best displays the hierarchy that answers clearly and concisely your questions.
Related Article: Stop Torturing Your Data and Other Tips for Revealing Real Data Insights
Know how your data and color transfer information
The success of a graph depends in part on whether it creates a cognitive load for the audience. Cognitive load refers to the amount of information the brain can process at any one time. So, you want to make sure that the graphic elements combine to tell the clearest story with the least amount of effort on the viewer’s part.
For example, bar charts and pie charts can show the composition of the data equally, but bar charts are better at showing unit differences. These differences are important to demonstrate the accuracy of the comparison. Instead of saying there was a 20% increase in organic traffic, for example, you need a bar chart showing a 20% increase. With just one look, the end user can easily absorb the change.
In the chart below, you can clearly see that there were fewer rear-wheel drive (r) vehicles compared to all-wheel drive (4) or front-wheel drive (f) vehicles.
Good visualization focuses on accuracy when referring to measurements. Heat maps can show gradient changes, but they can be a poor choice for resolution when an audience wants to understand the distinct numerical differences between elements. For example, if a change in temperature of a degree or two is significant to your subject, you need to choose a graph that highlights when this difference appears.
Color is another element to consider. Sticking to one color and using shadows to indicate visual discrimination reduces cognitive loads. Also consider accessibility concerns, such as users with colorlessness, when deciding your color scheme. The second color is acceptable to highlight a specific dimension so that it stands out against other dimensions in the bar graph. Two colors are ideal for graphs that show contrasting contrasts, such as a heat map. You often see this in correlation charts, like the graph below, to indicate the strength of the correlation for the notes.
But there are limits to the number of colors that can be assigned in some composition charts. Typically six to eight colors are a good playground to show meaningful difference across multiple dimensions or scales. More than that provides a lot of detail. The resulting visualization crowds graphic images together and makes it difficult to visualize the differences.
If more than eight different dimensions have to be shown with distinct colours, a grid layout is the best option. A wireframe is a diagram of nested rectangles displayed as a hierarchy according to the value of the data provided. The area of each rectangle corresponds to the numeric value of its data. Sizes make the scale of each data point clearly visible, with color scales providing greater distinction, all within a restricted viewing area.
In addition, advanced visualization platforms such as Tableau and Google Data Studio have options for querying subsets of data from data sources. This gives you additional options for colors and visuals to tell the story of your data.
Related Article: How to Effectively Use Google Analytics TreeMap Reports
Choose the visuals that fit your timeline or location
The next visualization option is to show how the data has evolved over time. Relationships graphs usually work well, such as line charts that can display a comparison over time, or regression charts of data changes over a specific time period. But you may have to view long periods of time to show an important, albeit slowly developing, trend.
This is where programming languages like R and Python can help. Libraries – scripts added to functionality – offer visualization options so the user can comment on graphs and create animations that show how the data changes over time. The data is often read in the program, and then mapped into visual graphs using the library. Python users have a choice of libraries, such as Matplotlib and Seaborn, while R users can access ggplot2, a library that relies on the grammar of the graphics concept to add or remove each graph element as a layer to provide customization options.
The advantage of these libraries is that you can create custom visualizations to suit your needs, using scripts that call data in real time through an API. This allows the graphs to stay updated with the latest information.
These are also useful for spatial visualizations such as geolocation graphs. The data is mapped to the location of interest, adding another consideration to displaying the information. The libraries for both Python and R offer options for visual maps and graph sets.
Ask how frequently graph updates are needed
Does the chart need to be updated on a regular basis to monitor continuous performance or is it necessary for a one-time analysis? The answer determines the type of workflow that works best.
Usually, real-time graphs are paired with cloud-based dashboards for data management and visualization. For example, in R programming, you can easily create a brilliant application, which is a simple web application that allows data, program results, and graphs to appear in a common digital environment. A brilliant app can be hosted as a dashboard that instantly updates visualizations as data is called up. Furthermore, you can also add HTML features such as buttons and sliders to allow your audience to adjust the presentation without touching the underlying data or code.
Ultimately, you should determine which reporting schedule best handles what your audience needs from the data. Doing so will highlight the steps needed to present your infographics and see what influences decisions. Sometimes there are technical reasons to adjust the schedule. People often prefer a static image or are limited to an image if the graph is a printed material. Assigning the raw data to the visuals raises the question of what is the access to the data sources needed to feed the graphs. If they are updated regularly, you need an easy way to update the data and associated annotations.
Related article: What is a dashboard? How business intelligence inspires growth
Some final tips on choosing good visualizations
In the end, choosing a good visualization will clarify your analyzes. As I mentioned in 10 Mistakes to Avoid When Rethinking Your Analytics Strategy, you want to avoid general questions that turn into a long, boring narrative about your data. This does not lead to meaningful conclusions about your marketing efforts.
If you have a lot of important material but you know your stakeholders don’t have a lot of time, you can put those visuals in an appendix so recipients can review the details when appropriate. You can discover more ground rules in my visualization post.
Choosing good images to tell a story puts the focus on your marketing analysis. Strong visualization will open discussions in your audience about the junk food that moves customer experiences — and your organization — forward.
Pierre Dubois is the founder of Zimana, a small business digital analysis consultancy. It reviews data from web analytics and social media dashboard solutions, then makes web development recommendations and actions that improve marketing strategy and business profitability.