The 5 steps of data visualization

The world is getting more complex by the day. People are expected to understand more in less time. This is not least due to the abundance of data that flows into our society every day. And it is not without reason that more and more data analysts are needed to immerse themselves in this data, looking for information that leads to new insights for specific issues. But it doesn't stop with just data analysis. Audiences often request that data is represented in a simpler way, so that they can get insights into a particular issue, problem or situation more quickly. And this is where data visualization comes into play. It is the bridge between data and design, and is about making decisions about what to show from a dataset and how to show it. At Clarify ?? we have developed a method with which we achieve proper data visualizations in 5 steps. This article explains all about it.

1. Information analysis

Every good data visualization starts with immersing yourself in the challenge: who are you making it for? What message needs to be conveyed? In which context does this take place? Then it's time to parse the data. This requires some knowledge of statistics. Time for a freshup!

Constants vs. variables

Data consists of constants and variables. Constants, as the name suggests, remain constant over time (e.g. the mains voltage in the Netherlands or the boiling point of water). Variables, however, can certainly change (e.g. temperature or speed). By detecting this difference, you can make choices about which data should be given more or less attention in your visualization.

Qualitative vs. quantitative variables

You can then distinguish qualitative and quantitative data amongst variables. Qualitative data say something about the quality of something (e.g. color or animal species), while quantitative data say something about the quantity of something (e.g. temperature or speed). This gives you guidance on how you can represent things in your visualization: in words or numbers, and whether or not supported with images.

Discrete vs. continuous quantitative variables

Amongst quantitative variables, discrete and continuous variables can be distinguished. Where continuous variables can take on any conceivable number, discrete ones can only do so in steps (e.g. number of employees within a company, after all, half employees are not possible ;) This also gives you an indication of how you can represent things: in absolute numbers, or maybe better to tally?

Nominal vs. ordinal measurement level

Finally, within data visualization, 2 different measurement levels for variables are possible: nominal and ordinal. While at a nominal level there is no specific order between values of variables (e.g. nationality or blood group), at an ordinal level there is certainly an arrangement (e.g. junior, medior, senior). For variables with an ordinal measurement level, you could decide to visualize a certain hierarchy, or place them one after the other in a certain reading direction.

At Clarify we ultimately make an information analysis of this, which often consists of an organized data document.

“Data visualization is the bridge between data and design, and is about making decisions about what to show from a dataset and how to show it.”

2. Structural diagram

After the information analysis, you take the first step towards a visual design, the structural diagram. This is the skeleton of the data visualization in which we do zoning to determine what goes where. The structural diagram provides substantive and visual structure and forms the basis for the subsequent steps.

A guideline for the structural diagram is the 3-stage model Angle, Framing, Focus. First you determine the angle of your data visualization (Angle). This is the most personal part: what do you think is important to show? What accent do you want to place? Then you think about how you want to convey the message (Framing). With which texts and which visuals? Finally, you determine which parts of the visualization you want to draw attention to (Focus). What should people really not miss? And what is supportive and can therefore be moved more to the background?

3. Data elements

Now that the content has been determined using a structural diagram, you can continue further detailing the data visualization. Using brainstorming techniques such as Mind Mapping and Brain Drawing, we ask ourselves how things can best be represented (Framing). In this phase we also further develop the storyline and determine the reading direction of the data visualization. These are important to guide your reader through the data step-by-step, so that it can sink in better and ultimately create more understanding.

In the data visualization toolbox we usually play with the following data elements:

Numbers

The most common element in data visualizations, especially suitable for continuous quantitative variables. In the right font and size you can display numbers in an attractive way in your data visualization.

Length

An alternative way to represent quantitative variables is by length. A bar chart is the most classic example of this, but in fact all types of lines with or without curves can be used to represent values. By using length as a visual element, different values can be compared more quickly.

Scale

In addition to length, quantitative variables can also be represented as areas. The higher the value of the variable, the larger the area in your data visualization. By using scale as a visual element, different values can be compared more quickly. Instead of areas, you can also scale icons to show the values, creating an even more attractive visualization.

Tally

Tallying is particularly suitable for discrete quantitative variables. This allows you to show different quantities between variables at a glance. And instead of dashes you can also use icons that indicate which variable is shown.

Percentages

If you are dealing with fractional parts, you usually express this in percentages. Just like numbers, you can display percentages attractively by choosing the right font and size. Percentages can also be displayed using charts such as pie charts.

Surfaces

A special form of percentages in data visualizations are surfaces. Just like with percentages, you show fractional parts, where all surfaces together form 100%. The great advantage of this representation is that it is immediately clear to the reader how the different values compare to each other.

Order

Simply arranging different values in order is also a way of data visualization, especially for variables with an ordinal level of measurement. By arranging them, the reader receives information about the hierarchy between the values.

Color

You can also make the values of variables clear using color coding. A legend is often needed to explain what each color means. Keep in mind that color is always a dominant factor in data visualizations. Therefore, carefully consider which data you want to focus on and which data may be pushed more to the background.

Icons

Icons can be widely used in data visualization, especially where no concrete numbers are available (such as with qualitative variables). Many of the aforementioned visual elements cannot be used in such cases, while icons here can directly represent or support the variables.

Networks

A special form of data visualization are networks. Rather than visualizing the values of variables, you visualize the relationships between entities (often constants). By representing these entities as dots or icons and drawing lines between them, you visualize how they relate to each other where the whole forms a network.

4. Visual design

In this phase you create the data visualization in the chosen visual style. While in the previous step you were busy with the Framing of your data visualization, you focus on the Focus aspect during the visual design. You determine where the focus points will be using the shape, size and color of your visual elements. This also allows you to control the reading direction of your data visualization.

Your customer's branding usually also comes into play in this phase. Are you tied to certain typography, colors and other corporate identity elements? Then this is the time to implement it, maintaining the balance between a functional, readable data visualization and an aesthetic end result that suits the customer.

"By adding interactivity, users can play with the data, allowing them to better absorb and understand the data."

5. Interactivity

In the case of an interactive data visualization, the design is then converted into front-end code. This includes all visible elements of the data visualization and ensures that the user can interact with them. The big advantage of adding interactivity is that the data can be shown in a more controlled way. For example, you can use toggles, radio buttons, checkboxes, dropdowns, tabs and search fields to give the user the choice of which data is shown. In fact, they all act as a filter on the data. This way of searching through data is called 'faceted search'. This allows the user to play with the data, so that the data is better absorbed and understood. So if you have the opportunity and resources to make your data visualization interactive, go for it! Of course at Clarify ?? we can also help you with this.

Happy data viz!