Motorcycle accidents, generalized linear models, prediction model, collector roads, traffic flow, length of road, access points, traffic speed
The phenomenon of traffic accidents in Indonesia is characterized by the high frequency of traffic accidents and the involvement of motorcycle. The involvement of this type of two-wheeled vehicles is the highest compared to other vehicle types. In addition to arterial roads, the frequency of traffic accidents in urban areas also often occurs on collector roads. This paper will be present the development of prediction model of motorcycle accident on collector roads in Surabaya, Indonesia. This study uses generalized linear model with a Poisson distribution and logarithmic link function approach, as well as the R software applications in the data processing. The location taken as a case study is urban roads. In this study, 69 sections of 120 collector roads in Surabaya are chosen. The result of this study indicates that the traffic volume, length of roads, number of access points, and the speed of traffic are significantly influential in describing a motorcycle accident on collector roads. These findings are expected to help the stakeholders in the field of traffic engineering to overcome the problems of traffic accidents on collector roads in Surabaya and other areas.