Visualize This: The FlowingData Guide to Design, Visualization, and Statistics

What Relationships to Look For

So far you looked at basic relationships with patterns in time and proportions. You learned about temporal trends, and compared proportions and percentages to see what’s the least and greatest and everything in between. The next step is to look for relationships between different variables. As something goes up, does another thing go down, and is it a causal or correlative relationship? The former is usually quite hard to prove quantitatively, which makes it even less likely you can prove it with a graphic. You can, however, easily show correlation, which can lead to a deeper more exploratory analysis.

You can also take a step back to look at the big picture, or the distribution of your data. Is it actually spaced out or is it clustered in between? Such comparisons can lead to stories about citizens of a country or how you compare to those around you. You can see how different countries compare to one another or general developmental can progress around the world, which can aid in decisions about where to provide aid.

You can also compare multiple distributions for an even wider view of your data. How has the makeup of a population changed over time? How has it stayed the same?

Most important, in the end, when you have all your graphics in front of you, ask what the results mean. Are they what you expected? Does anything surprise you?

This might seem abstract and hand-wavy, so now jump right into some concrete examples on how to look at relationships ...