A Time-Series Analysis of Traffic Crashes in New York City

Abstract

In New York City, traffic crashes are one of the main causes of fatalities in the city. This study presents a comprehensive time series analysis of road crashes in the city from 2013 to 2019. The crash data were collected, organized, and analyzed at different time levels: yearly, seasonally, monthly and hourly bases. Forecasting of the total number of crashes in the years 2020 to 2025 was conducted using the Box-Jenkins method based on the autoregressive integrated moving average (ARIMA) model. The model was statistically validated using a modified Box-Pierce (Ljung-Box) Chi-Square test. The proposed model was also used for backward prediction of the year 2019 to compare with actual observations. The predicted results showed a good agreement with the actual observed results. The results also showed a strong potential of having a reduction in the total number of crashes in the future.