Consider the usual situation where vehicles from a company regularly perform trips from a given location A to another location B. If all vehicles use the same routes, the trips are bound to have similar parameters, such as total traveled distance, trip duration, fuel consumption and so on. If one vehicle takes significantly longer to execute this trip, then maybe something went wrong, something worth looking into. Were there unusual traffic conditions? Maybe there was a flat tyre? Did the driver take a wrong turn somewhere?

This is what a Trip Outlier is all about: a trip from A to B that is *markedly different* from all other known trips from A to B; an exception. The difference can be detected according to a number of different measures.

For each of these measures, Frotcom Analytics will determine if the corresponding value lies outside of the expected range, which is calculated using a sample of previous trips between the same two locations (A and B) and in the same direction (from A to B). The expected range of values is determined through a robust statistical procedure.

Next: Trip Outlier Measures

## Comments