The DataHero Blog

The Art of Visualizing Data to Find Actionable Insights

February 26th, 2015

ACTIONABLE INSIGHTS

Charts, graphs, and other data visualizations, to be effective, should tell a story. The information relayed should give invaluable insights that will help the viewer make business decisions. Jon Steel, a leader in the account planning field, said the following of looking at and understanding data: “In the context of an advertising agency, the ability of planners to look at the same information as everyone else and see something different is invaluable. They need to be able to take information of all sorts, shuffle it around, and rearrange it in new patterns until something interesting emerges.” Not only will good data visualization convey actionable information, but it will help you to see things other people maybe can’t see.

 What can you do to get more actionable insights from your data? We’ve already talked elsewhere in our blog about comparing time ranges, looking at relationships and correlation, and looking at trends rather than at individual data points. Here we’ll talk about aesthetics, getting another set of eyes to look at the data, and why it’s a good idea to double-check your data.


Make it Look Good

XU7rR

Aesthetics are important when it comes to data visualization, but this doesn’t mean that the graphs and charts need to have a ton of colors and effects. Here, we can subscribe to the old adage that “less is more.” Consider the Darkhorse Analytics gif above. The graph starts with a lot of in-your-face fluff: backgrounds and borders and redundant labels and colors and special effects—bleh. Getting rid of all the extraneous effects makes for a very clear chart with easily discernible information. DataHero produces this type of chart, the type of chart that will cut through the clutter and let you see what really matters, plain and simple.

Less may be more, but that doesn’t mean you should completely forgo any effects. Play around with one or two effects to see what best represents your data or most helps the viewer understand the data. Color, for example, can make a huge impact in a positive way—but can also be overdone. Keep your color palette simple for the best impact. Consider these two pie charts. Where does your eye go as you look at them? Instead of using 8 different colors, use 4 different colors and 4 lighter versions of those colors. The viewer’s attention will be much more easily and naturally drawn throughout the charts this way. For more tips and tricks, read our post on color here.

Color Wheel Image

Have Someone Else Take a Look

Even if you’re pretty clear on what you’re seeing, get another set of eyes to take a look at your charts and graphs; one person can’t always see everything. Maybe you’re looking at your website’s traffic on an analytics platform and are upset because it seems lower than you expected. Someone else looking at the same graph might notice that you have filtered the traffic to only show referral or organic traffic instead of all traffic, or you have only a week view selected when you meant to view an entire month or even several months worth of data.

Interdepartmental communication may come into play here. Maybe you aren’t sure that event goals in your analytics are set up correctly. If the data doesn’t look right and fixing it might be over your head, reach out to IT or the DEV team to make sure things are set up correctly. You’ll get clean and clear insights as to what your data is saying.

Double-Check Your Data

Be skeptical with your data. Question what you’re seeing and look at it in as many different ways as possible to make sure you are understanding it correctly and interpreting it how someone else might see it. You don’t want to unintentionally mislead anyone, and you certainly don’t want to intentionally deceive. Consider this article presented by Forbes. The author presents several different circumstances that could result in misrepresentation of the data. These include creating bar charts without a zero, using evenly spaced tick marks to represent uneven intervals, creating graphs not drawn to scale and using area or volume for changes in one dimension. Be careful to look for these things when analyzing charts of data and take care to not replicate them when creating your own charts.

graph

Take a look at the graph above. After a minute or so of studying the graph, you will probably find that the bars start at $55 billion instead of at zero. This small detail actually changes significantly how you perceive the information. The second bar is twice as tall as the first bar which would indicate that it is twice as much, but the reality is that it is actually just an increase from $60 billion to $66.2 billion. The value in having a graph like this is that you can get information visually, and misrepresenting the data like this greatly devalues the graph and a main reason for having it.


There’s a lot you can do with visualizing data, but the real artistry comes in displaying it in such a way that brilliant, actionable insights emerge where they weren’t previously visible. DataHero works with your data to make it visually appealing, understandable, and easy to check for accuracy. Have DataHero help you create data visualizations like this and your data will speak to you.

By Paxton Gray

Create my Free DataHero Account

Get the fastest, easiest way to understand your data today.