Aside from the history and explanation of various data visualization types (data maps, time series, narrative graphics, etc.) I found that there were 5 major recurring themes in Chapter 1 of The Visual Display of Quantitative Information. Intuitiveness, effectiveness, efficiency, multivariation, and integrity. I more-or-less agreed with most of Tufte's recommendations, although there are a few personal revisions I'd like to make.
Intuitiveness
Tufte hit the nail on the head with this one. It's important to use common sense and adhere to the basic principles of whatever medium is being dealt with for that portion of the visualization. Example: if you're drawing a bar graph by hand, please use a ruler.
Effectiveness
Getting the message across. If the data or correlation is overly complicated, there's no real reason you shouldn't simplify it to increase receptiveness.
Efficiency
I felt that Tufte placed too much value on space and over-stressed minimizing the use of ink. Yes, it's important to try and reduce the amount of time it takes to communicate an idea, but there are times when ideas shouldn't be broken down any further. If the visualization is electronic, consider having optional, expandable areas of the visualization where viewers may have access to more information should they wish to pursue it.
Multivariation
I disagreed with Tufte on this one. "Graphical excellence", as he calls it, is not nearly always multivariate. True, there are times when more data is desirable, for comparisons or whatever other purpose you can imagine. However, there are also times when it's ideal to have as little variables as possible. The less variables involved, the simpler the data, the more straightforward the visualization, the easier the communication.
Integrity
A no-brainer, really. Unless it's the purpose of the visualization, it's unprofessional to fabricate data or skew graphical representations to help make a point. (Ever noticed how America appears huge on some maps and much smaller in others?)
