Search Yieldify

When data is not enough: why you need to learn the art of visualization

Data visualization is changing how brands can optimize their e-commerce channels in 2015. Every day, new streams of infographics populate the blogs we read, visualizing everything from spend trends to device usage to purchase geolocations.

But data visualization isn’t just a novel way to broadcast a data set. As Dr Johanna Kieniewicz, Head of Outreach and Engagement at the Institute of Physics points out, data visualization is a tool of discovery as well as communication.[1]

Business managers who use visual data discovery tools are 28% more likely to find timely information than peers who only use managed reporting and dashboards[2]

So what is the commercial application of data visualization? And why should your brand be using it to influence your online strategy?

Visualization helps tell the story of your data

Humans are visual creatures. We consume information more rapidly when demonstrated through diagrams than when expressed as text.

Big data visualizations are appealing and helpful when we have large data sets because they allow us to conceptualise by dimension; it is difficult to conceptualise data beyond three variables. Visualization helps us to intuitively build relationships between variables to get the full story from your data. 

Saurabh Johri, Data Scientist, Yieldify

The cumulative popularity of products over time. Visualization by Yieldify Data Science team

Visualization gives you better ownership of your brand’s data and enables you to understand the dynamics of your customers

At this moment, humans are better equipped to see visual patterns than computers.Machine learning algorithms are growing ever more advanced but their accuracy depends on having access to a body of images and words to train from. We are still some way off a computer being able to compete with the human eye and brain.

It’s not just about answers: visualization helps you ask the right questions

It’s easy to think that visualization is supposed to give you all the answers. Most people anticipate a visual will immediately reveal the source of a problem or, for example, that there is an obvious pattern of abandonment for a particular item.

In reality, visualization should help you ask more questions and better questions. Effective visualization helps push you towards finding better hypotheses. It should be a key stage in a constant iterative cycle of testing and refining.

Data insights: a visualization (Gregor Aisch) [3]

Data insights: a visualization (Gregor Aisch)[3]

Following this back-and-forth process, you can work out if the data is making it difficult to tell a particular story or whether it is telling a different story altogether.

Good visuals do not always mean good data

Whereas it is clear that data visualization has real benefits, better data and better visuals do not go hand in hand. We tend to immediately attach credibility to data visualization because displaying findings in this way makes them look objectively true. But it is possible to skew what you are trying to present. This is because there are many details that are hidden in a visual. If you were trying to make a point about data by using text rather than an image as the medium, it would be much easier to point out what you disagree with.

Be attuned to how that data is presented so you can prevent yourself from being misled by a bad visual. Take, for example, colour scheme: this can have a big impact on how the data presented is perceived. Often cosmetic choices can lend more credibility to the data than it deserves.

If for example, you highlighted an area on a map in red, it would scream out to you, when actually this would be a completely subjective choice made by the person who created that visual.

badgraph2

The colour would trigger mechanisms which humans are primed to react to. You should always be aware about the quality of the data that has gone into a visual.

In industry, data is siloed

Data visualization is often seen as a way of making your results look more quantifiable. In fact, they are not quantifiable in the same way that a robust scientific study would be – where others have access to that data so they can repeat the experiment.

“In industry, people can’t replicate what you’ve done to test its validity. Also, often what is seen in a scientific study ends up being glossed over in a commercial context in order to create a captivating visual. This is because you want to show something more direct to tell a clear story.” Saurabh Johri, Data Scientist, Yieldify

Be mindful that data visualization always has an editorial side.

Visualization enables data democratisation

Ultimately, your organisation is stronger when everyone gets it. Displaying data internally using visualization can excite, inspire new approaches and communicate insights that had before now been closed off to the rest of your team.

By building data visualization into a constant cycle of learning, new actions to improve your strategy will frequently become visible. These actions would have remained hidden possibilities if you hadn’t taken advantage of this game-changing practice.

 

[1] http://www.theguardian.com/technology/2014/feb/16/visualise-data-change-life-florence-nightingale
[2] http://data-informed.com/optimize-ecommerce-analytics-visualization/#sthash.bKsmje16.dpuf
[3] http://datajournalismhandbook.org/1.0/en/understanding_data_7.html