Despite the fact that nowadays there exist unprecedented opportunities to accurately monitor and analyze driving behaviour, the exact amount of the necessary driving data that should be collected for each driver in driving behaviour assessment is not determined yet. The objective of this paper is to quantify the need for driving data collection in driving performance assessment based on data collected through Smartphone devices from a sample of one hundred and seventy one (171) drivers that participated in the 7-months designed experiment. A statistical analysis was conducted to determine the period at which driving behaviour metrics (speed, mobile usage, harsh acceleration/braking events etc.) rate converges to a stable value. The impact of this methodology lies on the fact that it is essential for researchers nowadays. The impact of this methodology is significant since both small and big data samples lurk the risk of leading to doubtable results, by acquiring a sample either biased or computationally expensive to analyze.
|Tags||big data, driver behaviour, machine learning, naturalistic driving, statistical modelling, telematics|