The Art and Science of Data Visualization

In some companies, however, there is more inertia to using analytics, and getting potential users to improve their adoption of it. After assigning good talent to analytics (technologists, statisticians, business experts), giving them the right environment, patronage, budgets and tools, they find that somehow the efforts fail to go further in terms of genuine adoption of the results by the business. Assuming that the quality of output is good, there could be a number of other reasons why that output is not used to derive business benefit. One reason could of course be the culture of the organization, and the lack of adequate drive and interest in analytics from the top. But if that is in place, another reason could well be that the results are just not presented to the business in a manner that makes them understandable, credible and interesting.

The presentation of results is best done through visualization of the data. Statistical formulas and columns of numbers in spreadsheets may hold fascinating meaning for the data sciences team because they are already intimately familiar with the details, but to the lay business user they are nothing more than just lifeless numbers. They need to be made more easily understandable through the right data visualization technique, and brought to life with the right story to present them. There are a large number of proprietary and open source data visualization tools, but knowing how to use them effectively is both a science and an art.

Any presentation of ideas and data, no matter what the purpose is, is ultimately an exercise in selling. And this is why it is important to not just present the facts, but to make them interesting, relevant and credible. There are a variety of presentation formats to choose from. These are:

Simple data plots, line diagrams, bar charts and basic pie charts

For very small amounts of data, or for very simple charts that do not have many variables, MS Office Excel charts could be adequate, but are very basic and limited in terms of shareability, portability across device formats, online use, choice of data sources, embedding, etc. Alternatives that allow for more visually appealing charts and more developer flexibility include Tableau, Fusioncharts, Google Charts, Chart.Js, and many more. For higher data volumes, dygraphs is a popular alternative, available as a Javascript library.

More complex charts overlaid on a reference background

Sometimes, data needs to be plotted on a reference background, such as a map or a timeline. For map plots, Leaflet, InstantAtlas, Polymaps, CartoDB and Datawrapper offer a variety of capabilities and features. For plotting data on a timeline, Timeline.js is one of the most popular.

Infographics

Infographic presentations are probably the easiest to understand because they, more than any other technique, allow for impactful storytelling to be done using a combination, of facts in the form of text, numbers and images. There are tools such as Info.gram, Piktochart and others that offer an impressive number of templates and icons that allow blending of data with text. Achieving impactful results with these tools is only limited by the presentation creativity of those who use it.

For these reasons, even though the talent and expertise of the analytics team may be top notch in terms of the quality of their output and findings, their ability to choose the right presentation technique, and to use the right tools to highlight the key points while screening out the noise in the data may make a big difference in terms of earning the attention of the business. Given the range of options available today, it may be well worth the additional investment required to train them on effective presentation techniques and on presentation preparation as well.