From the historical dataset of a city, it is possible to not only just know why vehicle collisions in the past happened, but also predict where and how many collisions in a city would happen. In this project, the dataset of weather, holiday, and details of collisions in Ottawa from 2014 to 2017 would be organized, and divided into a series of graphs and models by different variables. These graphs and models helps statisticians to understand which variable(s) might greatly affect the probability of causing a collision. Furthermore, statisticians could summarize those situations that probably lead to vehicle collisions and prevent people from driving in similar situations, or predict the number and positions of vehicle collisions in a city.