Often the fastest graph to generate is not the simplest way to communicate meaning. For an analysis of trends over time in pharmacoeconomic graduate student interests, take a look at this figure before and then after modifications to improve visualization of survey results.
The first graph under-utilizes the ordering of years by assigning color categories to bars. Placing year on a continuous x-axis reduces complexity by one dimension. Directly labeling the lines, and deleting the traditional legend, facilitates easy interpretation of the current ranking and how the relative order has changed over time.
Do you have any favorite before and after data visualizations?
Special thanks to my fellow Survey and Evaluation Committee members at the ISPOR Student Council who designed and led this data analysis.