The Basics of Decoding Data Viz

Jan 17th, 2019 • 7 mins reading time

Reading or communicating data can sometimes feel daunting, especially when relying on visuals.

We can be afraid that it might be confusing or too difficult to understand. It’s easy to forget that the elements of ‘reading’ can sometimes be invisible and this is true when you think of body language, emotion and conveying expression.

We actually rely a lot on context to fill in the gaps. We can identify if something is happy or angry, looks right or wrong, or just look normal and we’re content with it. This ‘invisible’ aspect seems missing but in fact, can be decoded by understanding context.

How does ‘context’ relate to data visualisation you may ask? Although data viz cannot be as black and white as communicating with words, it can be argued that we read data visualisations mostly from context. Maybe this is the natural way to read charts?! We should trust the ability to use and allow context to open the doors to greater understanding, meaning and insight.

From our experience, there’s so much potential lost because of being safe by sticking to entry-level charting and assuming other people may not understand. Simple charting can only say so much. It’s one insight, one meaning and nothing more. The opportunity for greater insight can, in fact, be easily achieved, and you’ll be surprised that most people are willing and wanting to learn more.

By breaking down the elements and components of what we use to communicate data visually, let’s start by removing all labels, chart types and complexity to highlight what the fundamental reading components are.

The important to note that all forms of charting are somewhat familiar already, and the good news is that the majority of people already understand how to read basic charts or data viz.

We hope that this article sheds some light into decoding data viz and provides a stepping stone to make it more approachable and to help read more complex visualisations or maybe it’s a simple reminder of what you already know. This also certainly isn’t an ‘everything you should know’ type of article, but we’ve pulled out what we feel are the core visual principles when decoding data.

Reading Geometry

The word geometry maybe a fancy word for shapes but actually the breakdown of geometry is a powerful tool when we consider points, lines, angles, and surfaces. It’s only when you connect these attributes together you form a shape and in the data viz sense, form a story. Don’t forget, we also read the empty space just as much as solid filled geometry and it’s just as important too. We detect edges, completed contours etc. and all these things we already know and take for granted.

Sounds complicated? Let’s use some visuals to help explain.



If a surface area is empty we simply think of this as empty. So when applying a value to this context we give it 0% or nothing. If the surface area appears solid we can think of this area is now as complete or partially complete. The amount of surface area used we can fill up can represent a data value.

If we take this concept one step further and use the edge of a base shape to imply the total value, each segment of data plotted onto the surface area can add up to the total value eg. 100%.

Sometimes data might be distributed across a total value. An example of this might be cost distribution of services. We can, therefore, plot segments that represent the services across the whole surface area for comparison.


Lines are versatile because we can use them in a supportive role to overlay targets or averages but can also show trends and change-over-time (or other measurements eg. distance) in data. We can do this by plotting data values onto the line with the height indicating high and low values.


Proximity and scale

We just focused on the basic aspects of geometry in data viz but what about the empty space that surrounds it. We can use proximity to show the distance between one object from another and show different insights.

It could be as simple as allowing the dataset to do all the talking. If data is closer together, this generally means they are similar. Is that the expected result or are there outliers that need to be investigated?

Also, data points in close proximity could also convey a group of data for comparison. When comparing groups there may be an added level of complexity that may need colour to help.

We can use scale just as much as the surface area to represent data values and is useful for smaller data sets. It’s important to remember the purpose of the visualisation if that means, ‘this one is better than that one and by how much’ the visualisation has more than achieved the purpose having been backed up by data.

If the shapes are plotted by using both proximity (X and Y axis) and scale, the size of the shape will most likely represent a third value (Z).


Applying colour for meaning

There are a few ways to use colour, and this can range from grouping, categorisation or give positive or negative meaning to data.

Labelling data with colour makes the visualisations easier to scan. It’s completely normal to refer to a key to understand what the colours mean and is best to stick to few contrasting colours rather than similar colours.

Similar colours or shades of the same colour may indicate a range of high to low values. Usually the bolder the colour the higher the value and vice versa.

Visual indicators of colour can also be used as either an alerting mechanic for anomalies and most commonly in financial data, positive and negative. As most people (in western culture anyway) understand that green is good and red is bad, you just need to be cautious when applying a colour that it reflects it’s intended meaning. The data aspect comes in when you apply thresholds for example if it’s over 85% then it’s good. The colour green then serves as a powerful tool.

Now let’s add axis and labels

Now that we’ve covered a light overview of core visual principles in data viz, the axis and labels are utilised for both plotting data and telling what the data represents.

The axis and labels really dictate what level of detail and accuracy you want to go into when communicating your data. Is it about being precise or to make a point? Is for showing vasts amount of data or very little? Is it for art or for business-related decision making. Whatever the use case, data can be as creative or practical, complicated or simple as you want but the core principles of decoding data viz will always remain the same.


Aaron Nicholls

Founding partner of Studio Dot Info based in London. Tries to make sense of data and communicate it well through good design.


The Basics of Decoding Data Viz