The prevalence of road traffic collisions is a pressing issue both worldwide and within the United States. The consequences of these incidents are severe, resulting in loss of life, reduced productivity, and other socio-economic implications that demand immediate attention. To effectively address this problem, conducting an extensive literature review is crucial to identify the various causes of traffic collisions and the complex interdependencies between them. Addressing this challenge necessitates a targeted exploration of its multifaceted causes and their interrelations through an extensive literature review, incorporating the latest advancements in machine learning and deep learning techniques. However, the lack of a consensus on datasets and prediction techniques hinders the development of accurate, location-specific traffic collision predictions. By meticulously analyzing traffic collision factors and data sources and leveraging state-of-the-art ML and DL approaches, this paper endeavors to forge a pathway toward developing precise, location-adapted predictions for traffic collisions, thereby contributing significantly to the discourse on long-term preventative strategies.