Road Traffic Accident Trends, Contributing Factors, and Predictive Analytics in India: A Data-Driven Analysis (2021–2023)
TL;DR
Official crash statistics from India's Ministry of Road Transport and Highways for 2021–2023 with recent peer-reviewed literature on machine-learning-based crash-severity prediction show measurable gains in severity-prediction accuracy, supporting more targeted interventions.
Abstract
Road traffic crashes remain a major public-safety challenge in India, which records among the highest absolute crash and fatality counts of any country. This paper synthesises official crash statistics from India's Ministry of Road Transport and Highways (MoRTH) for 2021–2023 with recent peer-reviewed literature on machine-learning-based crash-severity prediction. National accidents rose from 412,432 (2021) to 480,583 (2023) and fatalities from 153,972 to 172,890. Overspeeding, non-use of helmets/seatbelts, and concentration of crashes on highways are the dominant factors; two-wheeler riders and pedestrians bear the largest fatality share. Recent ML/deep-learning models show measurable gains in severity-prediction accuracy, supporting more targeted interventions. Engineering, enforcement, and data-infrastructure recommendations are discussed. Keywords: road traffic accidents; road safety; crash severity prediction; machine learning; India; MoRTH