So far most of my blogs have been on “just for fun” applications, largely because I can’t go into specifics on business applications due to their confidentiality. But it’s the business world that’s really driving acceleration in the field, so I will try to redress the balance here in describing (in suitably broad terms) how visualisation is helping businesses make better decisions. This is the first in a 3-part special to illustrate the topic: Part 1 “What is visualisation?”; Part 2 “How is visualisation used in real life to improve business decisions?”; and Part 3 “What new types of chart are there to tell my data’s story?”
So what is visualisation?
By definition, it is the presentation of data in a pictorial or graphical format – but we’ve been doing that for years, the bar chart isn’t a new invention, I hear you say! So why has it now become a field in its own right? Data visualisation has evolved because there’s so much more data to convey, but the old adage “a picture is worth a thousand words” is just as valid today. Data displayed graphically is easier to understand and provides a way to show potential relationships which are not as obvious in tabular form. The primary goal is to communicate information clearly and efficiently, enabling decision-makers to better grasp difficult concepts or identify new patterns.
The art comes from presenting data in the way that is most intuitive to understand – which may be a bar chart, or else something new. For example, you wouldn’t expect to see a scatter plot used to show different regions’ total sales, a bar chart would be more intuitive – but why is that? When you’ve been working with data for long enough, the chart selection is almost instinctive, but what is the science behind the way the human eye takes in information from a certain type of chart that makes it appropriate? Understanding how humans see and organise information is critical to effectively communicating data and leads to more intuitive designs.
The eye is able to readily distinguish differences in line length, shape orientation, and colour without much effort (it uses “pre-attentive processing”), but shapes themselves are less straightforward. For example your eye finds it hard to identify the number of times the digit 5 appears in a series of numbers; but if that digit is different in size, orientation, or colour, it would be much quicker by making use of this pre-attentive processing. Effective graphics take advantage of this. For example, since humans can easily process differences in line length but not area, it is often more effective to use a bar chart rather than pie charts to show comparison. More on this in Part 3 of this series, looking at the different types of charts available and their best uses.
The field of visualisation is also about sifting a complicated concept down to its simplest form. This can involve treading a fine line between not distorting what the data has to say, and also not making the user suffer from information-overload. The best visualisations encourage exploration at different levels: an easy-to-spot overview message, with finer levels of insight if you look further.
Characteristics of Effective Visualisations:
- Impact, to drive your message home: less is more! Some charts try to pack in all the information available and end up with the user not knowing where to look. Ask yourself, or preferably someone else, is it clear where to start?
- Easy to understand, to include a wide audience: users who have only previously been used to bar charts should still be able to understand the chart and find the main insight that you’re demonstrating. Ask yourself, is it clear what the chart is intended to say? What questions might your audience have?
- Fun to look at, to engage the reader: whilst the above might tempt you to stick with bar charts, this guideline encourages the chart to be a little more playful. If the reader has been encouraged to explore the insight through your chart (without working too hard!) they are more likely to want to know more and be engaged and interested, and enjoy the experience.
- Audience-oriented, particularly if they’re decision-makers. Effective visualisations help users with their reasoning so the display of the information should be centrally informed by how decision-makers themselves will use it. Ask yourself what question the data is expected to answer?
Having said all that, that’s still not the extent of the value of visualisation. When people say “visualisation” in industry, they are usually referring to “interactive visualisation”, which adds an extra dimension of value – much more than just a pretty face!
What is interactive visualisation?
Frustratingly, interactive visualisation has sometimes been likened to a souped-up pivot table – which doesn’t do it justice at all, but is perhaps a good place to start for those of you who know Excel.
The similarity is that, like a pivot table, a large set of data can be summarised by category to give, say, average sales margin or profitability, broken down by region or product. Any performance metric can (should) be displayed in a more visual way than a table though: for example a gauge or a heatmap brings a table of data to life (more on chart types in Part 3).
Apart from just being able to see the results more clearly, with an interactive visualisation you can also investigate by drilling down to get an instant refresh of that metric. This lets you spot regions or product types that are contributing to a low performance for example. In the above, you might want to see why the result on the gauge is only average, you think it might be because a particular region in the West is under-performing, so you click into it:
You’ll notice all the charts update: the average margin chart has stayed coincidentally very similar because the performance of the region selected is about in line with the country as a whole, but the bar chart on the right has changed to reflect the product mix of the region selected – there are only 4 products here, and they have a slightly different profitability range. It’s not a big difference though, so it’s probably not this region causing the problem. Instead, you now think it could be a particular product type driving the profitability down (that much is clear from the bar chart!): is this product type’s poor performance emanating from a particular region? You go back to the overview and click into the lowest product type to investigate:
Again, the charts refresh instantly – this time to reveal a big change in both the average margin gauge (as expected) but also the regional breakdown for that product type. The selected product type is causing an issue for almost every region. This informs how you might need to change your strategy, in particular to focus on this identified under-performing product type.
You might already be seeing the potential from this simple example, as to how visualisation could add huge value to your own work, as well as that of big businesses. I’ll describe how in more detail in the next blog!
NEXT UP IN THIS 3-PART SPECIAL: HOW IS VISUALISATION USED IN REAL-LIFE TO IMPROVE BUSINESS DECISIONS? Visualisation in Business – Part 2