Now, let me summarize what we have seen during this week.
The main idea behind this module is the idea of channel effectiveness.
What I described at the beginning of this module is the idea that every visualization
can be described as the mapping between
some data properties and some visual channel properties.
And in general, the concept of visual channel is a very,
very relevant one in visualization.
And we discussed throughout the week of different properties of channels,
and properties that describe the effectiveness of a channel.
We also saw that effectiveness is in a way multifaceted.
It's a concept that is derived from many different properties.
So we saw that channel effectiveness depends on accuracy,
on disriminability, on salience,
on separability, and lastly, on grouping.
Let me very quickly give you a set of take home messages
for each one of these channel properties that we saw during this week of work.
Accuracy, the main thing to remember about accuracy is that you may want
to prioritize channels that are more accurate than others.
We saw that there is a ranking, whenever possible,
it's better to use channels that are more highly ranked than the others.
The second one was discriminability.
What is the main take home message of this disciriminability?
I think it's the idea that you have to remember that you
can't really perceive too many distinct values in every given channel that you use.
And typically, the number of values that you can distinguish is between five or seven.
So, this is particularly important with color,
but it's true with any other channel.
Don't expect your viewers to be able to distinguish more
than a handful of different values for each channel that you use.
The next one is salience.
The main concept to keep in mind about salience is that you have to be mindful of
how you are directing the viewers' attention in your visualization.
And also, the different channels are also competing one to another.
And if you have too many channels and too many values,
it's going to be really,
really hard for your viewers to focus on anything that they are interested in.
So you should never overload your graphics with too many channels and too many values.
Next one is separability,
and this is once again about the interaction of multiple channels.
Channels interact in many different ways.
But one important concept that we saw
related to separability is that some channels are perceived as integral,
whereas some channels are separable.
And you want to use separable channels when you want your viewer to be able to
attend to the information conveyed by one channel independently from the other.
Whereas, you may want to use integral channels when you want
the viewer to be able to perceive the combination of
these two attributes in a holistic manner.
And lastly, we saw the laws of grouping.
And the main take home message here is that,
you have to be mindful of how certain visual properties give a sense of grouping.
So objects get grouped together when they are arranged in a certain way.
And also related to that,
some grouping methods are stronger than others.
Enrico BertiniAssociate Professor
Cristian FelixPhD. Candidate
