A Simple and Explainable Model for Park-and-Ride Car Park Occupancy Prediction

TL;DR

This work introduces a simple, explainable occupancy model for park-and-ride facilities based on truncated-normal arrivals and departures, enabling accurate occupancy prediction and estimation of unmet demand using only aggregate data. The core TN model captures circadian occupancy patterns, while the extended TNL variant estimates capacity-threshold moments and overflow, informing capacity-planning decisions. Validation on Barcelona-area car parks demonstrates robust predictive and nowcasting performance with low errors on weekdays and clear utility for policy analysis. The approach emphasizes interpretability, low data requirements, and scalability for urban planning of park-and-ride infrastructure.

Abstract

In a scenario of growing usage of park-and-ride facilities, understanding and predicting car park occupancy is becoming increasingly important. This study presents a model that effectively captures the occupancy patterns of park-and-ride car parks for commuters using truncated normal distributions for vehicle arrival and departure times. The objective is to develop a predictive model with minimal parameters corresponding to commuter behaviour, enabling the estimation of parking saturation and unfulfilled demand. The proposed model successfully identifies the regular, periodic nature of commuter parking behaviour, where vehicles arrive in the morning and depart in the afternoon. It operates using aggregate data, eliminating the need for individual tracking of arrivals and departures. The model's predictive and now-casting capabilities are demonstrated through real-world data from car parks in the Barcelona Metropolitan Area. A simple model extension furthermore enables the prediction of when a car park will reach its occupancy limit and estimates the additional spaces required to accommodate such excess demand. Thus, beyond forecasting, the model serves as a valuable tool for evaluating interventions, such as expanding parking capacity, to optimize park-and-ride facilities.

Figures (16)

• Figure 1: Locations of the park-and-ride car parks analysed, all of them in the vicinity of Barcelona (Catalonia). Two of the locations, Martorell and San Quirze, have not been used due to errors in the data collection process. Scale indicates a 5km distance.
• Figure 2: Example of input data from the Vilanova car park, days with grey background have been removed from the analysis due to sensor failure (7th to 9th of February), COVID-19 induced lockdown (days after March 15th) or reduced activity due to Christmas holidays (1st to 3rd and 6th of January)
• Figure 3: Average weekly (top) and daily activity cycles (bottom) of the Vilanova car park. The daily average cycles are very similar from Monday to Thursday.
• Figure 4: Histogram of the observed maximal occupancies for the days between 01-01-2020 and 01-04-2020 aggregated by weekdays (orange), Fridays (blue) and weekends (green) for the eight car parks used in this study (bin sizes $= 9$ in all panels). Vertical red lines at the right of the panels indicate the maximal capacities of the car parks. The maximum capacity was reached in SantSadurni (19 times), SantBoi (39 times) and in QuatreCamins (45 times).
• Figure 5: Example of the TN model with data from the Vilanova car park. (top) probability density function (PDF) of car arrival (blue) and departure times (red), (bottom) corresponding CDFs (dashed lines) and car park occupancy curve (red thick solid line).