According to the single axiom of visualization, a figure should be judged according to its goals. Lets discuss it a little bit more. There are three consequences:
1. A figure may work amazingly in some circumstances but fail miserably in another. Take a look at this wonderful barplot:
It is perfect for media usage: it grabs attentions and efficiently delivers the emotional message. But it would not fit in a scientific publication. There are only 4 number in this figure! And only 2 values contain any information: the ratios between the deaths and expenses. One should newer use a picture to present 2 values, one just write them down in a plain text!
2. It may not be wise to look for the examples of good visualizations in media. These people may have pretty different goals from what you are looking for.
2a. Actually, it may not be wise to look for such examples even in scientific literature. Usually a visualization is made according to traditions and default settings in matlab/R. No aims or goals are considered.
3. There are no better judge that yourself and common sense.