Prediction of Car Parking Occupancy in Urban Areas Using Geostatistics

Public on-street car parking is an important shared resource of a city infrastructure with a significant impact on traffic. This paper proposes a geostatistical model aimed to predict parking occupancy rates for different periods of the day. In the study case, the occupancy representation considers the georeferenced position of spots for a particular area of Los Angeles (USA). Different models are compared and their parameters are estimated using the available dataset of the parking area. The final model is chosen to generate a kriging map that helps to understand and predict the occupancy rates. The end goal is to open doors for modeling and predicting urban phenomenons with Geostatistics to help with planning public parking policies in high density urban areas.