I have to provide choropleth maps for for Energy savings, Emission savings, Cost Savings related to different administrative boundaries such as provinces and states, but I don't know which classification is more appropriate because my data for each province is separated from other parts and there is no link between the numbers. Any suggestion for how should I classify? I don't want to use equal intervals because the minimum is 20 and the maximum is 400 and there are few over 300 figures so equal intervals represent my data not in a good way, but about other methods such as jenks or deviation or quantile I have no idea!
I've been experiencing the same kind of problem. The solution that I worked out at the time was using the 'Histogram' option in Qgis. In this option you can manually determine the classification of your graduated map. You can slide the vertical line to the appropriate value.
It's actually a very handy tool to use, since you can also see how many features will be captured in one class. This way you will be able to represent your data in the most desired way. (I Hope!)