Brian C. Hughes
CS 465 HW#3
9/16/08
Here's a picture that shows the general idea of my email visualization:
The idea is that emoticons represent the general trend of the email. In much the same way that spam
filters check for certain words and phrases, we could guess how the user felt based on the number of
words that fit into an “excited” word list or a “angry” filter. Spam has it's own icon as well (I thought
the devil was appropriate). For example, if you see a string of angry faces, you can safely assume a
flame war is going on, or perhaps a political debate gone bad.
Time is represented by the size and position of the icons. The newer mail messages are larger and
further to the outside, while older emails are smaller and near the center.
Chains of messages and replies are indicated by connected strings of icons. The color of the strings
represents the email address of the person you are talking to, with each address having a unique,
randomly generated color. I didn't think that spam should be considered unique, so I didn't include a
colored string for them.
You'd be able to estimate at a glance how much of the mail you get is spam, and watch the trend over
time. You would also be able to get the gist of an email chain with a particular user, like noticing that
conversations with a certain person are always depressing. I'm not entirely sure how useful this
visualization would be, but I certainly think it would be fun.