Discovering Puerto Rico Roads Profile Influencing Traffic Fatalities
The infrastructure of Puerto Rico (PR) was granted a grade point average of "D-" in 2019. Transportation crashes may be influenced, among others, by environmental or infrastructure hazards. Before creating and implementing the first PR Strategic Highway Safety Plan (SHSP), the island reported more than 200,000 traffic crashes annually between 2007-2013. They resulted in a staggering 360 fatalities and 5,200 seriously injured individuals. Road fatalities dropped by 20.72%, and serious injuries decreased by 36.93% throughout the SHSP 2014-2018 implementation. Despite this progress, pedestrian traffic fatalities continue to be a major worry. Is it possible to predict injury severity or crash circumstances for a pedestrian vis-à-vis specific roads that support the Puerto Rico Highway and Transportation Authority (PRHTA) safety improvement efforts? Authors in Indonesia have used the Naive Bayes algorithm to predict crash severity by analyzing contributing factors. This research aims to establish a PR road profile with potential high-severity crashes. Data mining and predictive analytics helped to forecast crash severity and its associated road location. Outputs have shown tree (standing only), live animal, wall, utility pole/light support, fell/jumped from the vehicle, other non-collision, and ditch as the first crash injury-or-damage producing event. With 88.1% likelihood of suffering a fatal injury, especially in 1-2 (Aguada and Aguadilla-Int. PR 459); CR-3 (Canóvanas, Carolina, Guayama (Int. Ancha de Blondet St. and Int. PR 54)); and 1-52 (Caguas, Cayey, Juana Díaz, Salinas, Santa Isabel). PRHTA could consider the findings for focusing funds' investments on safety improvement measures.
Authors
- Yesenia Cruz Cantillo
- Carlos González Oquendo
Publication Details
- Published
- 2024-10-01
- DOI
- https://doi.org/10.21872/2024iise_5130
- Primary Topic
- Traffic Prediction and Management Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00