Affinity Mapping
Discover order in complex data by grouping it according to its deeper connections.
it’s not effective to approach complexity with tools designed to manage the merely complicated; this pattern looks at one approach that can help.
“…the complexity of our world has far outstripped any ready-made theories or hypotheses, and a priori assumptions and wishful thinking are useless” — Jiro Kawakita, 1991
-
This process is also called the KJ Method after its inventor, Jiro Kawakita.
-
Affinity mapping is a process where you take a large amount of information and try to put it into an order of some kind. It’s not just a process of assigning a category to ideas; instead, you look for connections between bits of information and they are related that way.
-
In practice, this looks like a large table covered with all the information being organized in printed form. It can include anything that can be printed: photos, drawings, maps, timelines, and more as well as writing, but for simplicity, we’ll call them cards. As you go over the various bits of information, you place cards closer to other cards when you notice a connection of some kind.
“With masses of data spread about on my desk, I had been racking my brains to find some way to integrate them when I suddenly realised that depending on the spatial arrangement of the cards, you can see new meaning in them and find ways to systemise the data. That was the first realisation that led to the creation of the KJ Method.” (Kawakita, Cited in Scupin)
-
The process uses abductive reasoning to group the cards. In this context, you look at two cards and try to come up with a hypothesis that would connect them. For example, a card about a bird and another about a butterfly could be connected on the basis of their ability to fly or perhaps on the fact that they both migrate. As cards are clustered, an order emerges.
-
This pattern language was built with the help of affinity mapping.
Therefore:
Discover the inherent order in a large, complex set of ideas by systematically clustering them according to affinities they have with each other. Resist creating simple categories, and look for deeper connections between ideas.
Understanding and communicating complexity is itself a Design process, and is best done iteratively