This study was conducted to investigate and model the high traffic collision frequencies in the City of Edmonton, Canada. Consistent collision spikes were observed on Fridays compared to the other days of the week. The first Negative Binomial model was formulated to establish a relation between the collision frequency and the independent variables. The second Multinomial logistic regression model was formulated to examine the probability of age categories and gender involved in collision for each day of week considering collision has happened. The proposed collision prediction models were found good. They could provide a realistic estimate of expected collision frequency and properties of collision for a particular day as a function of number of hours of daylight, number of hours of snowfall, visibility, age and gender. It is hoped that predicted collision frequency will help the decision maker to quantify traffic safety of Edmonton and improve the scenario. / Transportation Engineering

No description available.