Dynamics of execution in vehicles tracking from historical events

摘要

Efficiency and accuracy are key parameters in the planning and execution of transport plans. Often drivers are not able to fulfill the planned schedule due to several uncertainties that can not be predicted. This problem has been tackled based on the empiric knowledge and experience of the people involved. Now with the increasing amount of tracking devices and new computational power and storage capacities, it is possible to apply data-mining and machine-learning processes to big amounts of historical data, allowing to explore the hidden relationships related with efficiency and accuracy trips results. This article presents an exploratory analysis of the relationship between scheduling and execution of land transport trips that are tracked through real-time monitoring in Colombia. A data mining techniques exploration is used in the analysis of historical data obtained from vehicles and drivers. The comparison of the trip planned schedules and its real execution, that is extremely volatile, evidences the keys variables that are relevant in the decision-making process related to the transport planning problem.