Doing data visualizations correctly takes careful consideration. Incorrectly visualizing something can be misleading, embarassing, and even damaging to reputations. In order to do it correctly, it can often be useful to think about the visualization from several different angles before settling on the final version. Looking at good examples of data visualization is certainly a great way to learn, but equal value can be found in examining visualizations that didn’t work so well.
I recently started a Tumblr for collecting examples of bad visualizations. The examples are often funny, but #WTFViz is not intended solely to be humorous. The examples are also there as educational material, showing what not to do.
The data going into the visualization is the best place to start when selecting what visualization to use. There are subtle but complex concepts contained in data, and those need to be reflected accurately... keep reading