Feasibility analysis of construction of predictive models for road accidents in the city of La Rioja

Abstract

This project, running from December 2024 to December 2025, aims to assess the feasibility of developing predictive models to enhance preventive road safety. To achieve this, historical and real-time data from monitoring systems will be used, considering dynamic variables such as weather, temperature, date, day of the week, time, and the occurrence of relevant events, among others.

 An exploratory-descriptive methodology will be implemented, beginning with an analysis of the resources currently in operation and those planned for incorporation during the 2024-2025 period, according to the Government of the Province of La Rioja. In particular, the infrastructure and operational capacity of the 911 system, under the Ministry of Security, Justice, and Human Resources, will be examined.

 In the initial phase of the study, the geographic distribution of operational surveillance cameras will be analyzed, evaluating their technological features and the data they can generate (such as image resolution, infrared vision, and sound capture). Based on this information, the most suitable predictive model will be determined, whether it needs to be developed, combined, or activated if an existing model is already applicable.

 Preliminary conclusions will be drawn from the analysis of data and the predictive models applied. Subsequently, analytical methods will be applied to case studies to design a control dashboard that identifies high-risk areas within the city. This dashboard will strengthen preventive road safety by providing key information for decision-making.

 Data analysis will rely on univariate and multivariate descriptive statistical tools, utilizing specialized software such as InfoStat and R, as well as any other necessary technological resources to validate or refute the study hypothesis. Additionally, road safety statistics will be essential to measure the problem, identify high-risk segments, and analyze the evolution of trends over time.