Techniques involving modeling and analysis of operational metrics in transportation industries are disclosed. Representative methods involve defining a set of fuel consumption performance factors having a potential to impact vehicle fuel economy. Fuel consumption performance factors are received, and categorical variables are identified from at least a subset of the fuel consumption performance factors to create one or more categorical groupings. The fuel consumption performance factors associated with at least one of the categorical groupings are processed to identify a hypothetical maximum fuel economy or “baseline” for the at least one of the categorical groupings, and to identify fuel economy deduction coefficients based on the fuel consumption performance factors that adversely impact the hypothetical maximum fuel economy. Models are created based on the hypothetical maximum fuel economy and the fuel economy deduction coefficients to predict fuel economy for vehicles in the at least one of the categorical groupings.