This chapter explores the complex interplay between climate change and the accuracy of traffic flow predictions, focusing on the crucial use of geospatial data analysis. The potential effects of extreme weather events, such as heavy precipitation, storms, and heat waves, on traffic patterns should be considered to improve the robustness of traffic management systems. In this study, the authors demonstrate the effectiveness of geospatial data analysis in considering climatic and environmental variables to improve the accuracy of traffic flow forecasts. By integrating data into predictive models, we provide tangible evidence of the impacts of climate change on urban traffic patterns. The results obtained from data and simulations on machine learning models such as Lasso regression, random forest, XGboost and LTSM gave us very good results. prediction performance on the random forest with a correlation coefficient of 0.94; an RMSE of 265 and a MAE of 279 thus demonstrating its effectiveness for predicting traffic flow.
TopIntroduction
Rapid urbanization in underdeveloped countries such as Cameroon, India, Niger, Angola and Thailand, combined with the effects of climate change and changing modes of transport, has led to significant challenges in road traffic management (United Nations, n.d.). Accurately predicting traffic flow is essential to ensure smooth traffic flow, reduce congestion and improve transportation infrastructure planning. However, climate change has introduced additional complexity to this task, requiring a thorough understanding of its interaction with traffic flow forecasts. Learning, a branch of machine learning, is used to estimate corn yield forecasts in Cameroon using geospatial climate data. Geospatial climate data includes parameters such as temperature, precipitation, wind speed, and sun exposure, which are key factors influencing crop growth and yield (The TIME, n.d.). This study aims to predict corn yield in Cameroon using geospatial climatic data using two methods, LSTM and GRU, to determine which is the appropriate model for a good prediction of corn yield using temperature as geospatial climatic parameters, wind speed, precipitation, and solar radiation. Our study focuses on comprehensively exploring the complex interplay between climate change and the accuracy of traffic flow forecasts, emphasizing the crucial use of geospatial data analysis.
The underlying problem lies in understanding how climate variations affect traffic patterns, which challenges existing forecasting systems. Extreme weather events, such as heavy precipitation, storms, and heat waves, can significantly affect traffic patterns, requiring adequate consideration of traffic management tools. The complexities associated with traffic forecasting arise from the nature of the traffic domain, which includes the constraints imposed by the physical traffic infrastructure, such as road network capacities, traffic regulations and management policies, and the behavior of individual agents (road users), as well as exogenous factors, such as calendar (i.e. traffic time). day, day of the week, etc.), weather, accidents and incidents, events, and road closures, to name a few (Essien et al., 2019). Traffic data science has evolved by expanding many data sources to develop predictive models. Early studies investigated the importance of geospatial weather data on traffic flow parameters influencing driving behavior, travel demand, travel mode, road safety, and traffic flow characteristics. Additionally, research has shown over the years that precipitation reduces traffic capacity and operating speeds, thereby increasing traffic congestion and loss of road network productivity.
For example, the authors (Essien et al., 2018) reported that precipitation intensity affected urban traffic characteristics by reducing traffic speed by 4-9%. In contrast, rush hour traffic jams correlate significantly with the temperature intensity. Despite the importance of weather as a traffic indicator, most traffic forecast models used in ITS assume weather conditions are clear, thus missing important sources of environmental data that could allow a more precise assessment of the state of the traffic network. Despite the importance of weather as a traffic indicator, most traffic forecast models used in ITS assume clear weather conditions, thereby missing important sources of environmental data that could enable a more precise assessment of the state of the traffic network (Agarwal, 2005). Geospatial data analysis has proven to be a practical approach to improve the accuracy of traffic flow forecasts by considering climate variables. Integrating geospatial data such as weather data and environmental characteristics into predictive models makes it possible to capture the complex interactions between climate change and urban traffic flows. Therefore, this study aims to demonstrate the effectiveness of geospatial data analysis in considering climatic and environmental variables to improve the accuracy of traffic flow forecasts. By providing tangible evidence of the impacts of climate change on urban traffic patterns, this research will contribute to improving traffic management systems in underdeveloped countries.