# Relative or absolute colors on a map

I am designing something that will use a colored map - with red (worst), yellow (middle) and green (best) for separate regions that would be better or worse for an individual to visit.

These depend on several user selectable settings. Considering that, currently I have this set to an absolute scale - if a user's value is below a certain threshold, he'll get a color for a lower value. So for instance if his score was, say, 7 out of 10, the region in question would appear yellow - even if that is the current highest score on the map.

The other way I am thinking about doing this is as a relative value - I will put all of the calculated values into an array and go through the first 1/3, and then the second 1/3, etc. This would ensure that 1/3 of all regions are green, 1/3 of all regions are yellow, and 1/3 of all regions are red.

Which is preferable to user experience and user configuration? What is less confusing to the user? My guess is that relative would probably actually be easier for a user to understand - it conveys more information that is relevant to the user - "Oh, those regions are the best on the map for me". Is this analysis correct?

TL;DR

It all depends on what you want to show your users and the message you are trying to send with the data.

Long version:

So what you describe are two methods of visualizing a map called "Equal Interval" (absolute) and "Quantile" (relative).

The Equal Interval Method makes sure that the same number of values are in each classification set. So in your case with 7 of 10, that would mean that even though it's the highest value on the map, it still falls in the yellow range (which could be between 6 and 12)

The Quantile Method make sure (as you described) that there is an equal number represented in each scale interval. So 1/3 of your map would be green, 1/3 yellow, and 1/3 red.

Here is an example of both methods:

So to get to your question(s)

Which is preferable to user experience and user configuration?

This honestly depends on you. In this situation, you're the map maker and you get to show the user what you want them to see.

Let's take a look at a set of maps I made a while back. These are four of the more major types of classification methods used in common maps. You have Quantile, Equal Interval, Standard Deviation, and Manual classification. There are a few others, but I'm not going to discuss those.

Quantile:

Equal Interval:

Standard Deviation:

Manual:

Anyways, these four maps all show a very different story with the same data (Obesity rates provided by the Census. All values are in percentages). So with that in mind, you need to ask yourself "What am I trying to show my users? Do I want them to be worried about how much red is on the map or should I try to show more green to make them feel ok about the data?"

What is less confusing to the user? My guess is that relative would probably actually be easier for a user to understand - it conveys more information that is relevant to the user - "Oh, those regions are the best on the map for me". Is this analysis correct?

Again, it goes back to what you want your user to see. There isn't a definitive method that is better than the other. However, if you want to get the full details on all the various methods to help you choose what is best for your users to see, I suggest you check out this article on DirectionsMag. They do a really good job of describing the different methods available for map classification.

• Wow, that was a really in-depth answer that explained this to me. Thank you! I'm still relatively new to doing a ton of data visualization. Apr 26, 2015 at 18:34
• Well, I was trained as a cartographer, so I know my stuff :). If you want more explanation on anything just ask!
– BDD
Apr 26, 2015 at 18:46