Development of a daily duty cycle based on real world driving behavior and parking times is a critical requirement in the optimal design of power-train components of a plug-in vehicle. Standard driving cycles cannot completely emulate the real-world power demand of a vehicle and its downtimes in particular. To address these shortcomings, a large database of one year of measured data collected from a fleet of 76 cars in the city of Winnipeg, MB, Canada is obtained and is then used to develop a new duty cycle. This paper describes a methodology for statistical analysis of the fleet data, including while a vehicle is parked. Due to the intrinsic differences in vehicle usage profiles during weekdays and weekends, two 24-h duty cycles with suitable windows of opportunity for charging are developed for weekday and weekend driving patterns. The uniqueness of the proposed statistical methodology and the resulting duty cycles contribute to addressing the present shortcomings of standard driving cycles.