Since the 2013 publication from our lab showing that agrochemical odor pollution has negative impacts on bumblebee foraging behavior we had been working on how to quantify odors, in order to be able to quantify odor-pollution. This was a non-trivial problem, as most methods of odor representation are statistical in nature, which means that changes the stimulus set (i.e. the odors you are looking at) changes the quantitative relationship between stimuli. Seven years, multiple experimental paradigms, and many many hours of work in the lab resulted in this publication. Here we demonstrate the efficacy of a multidimensional space that represents the stimulus energy of complex odor blends based on their functional group and carbon characteristics. This computational method is effective at both describing and predicting bumblebee behavior in an associative odor learning task.