Trucks' illegal parking results from parking shortage and policies that do not consider the parking requirements of trucks. Understanding trucks' parking behaviour is needed for the design of adequate infrastructure and the development of effective policies. For truck parking, data is limited, incomplete, and of variable quality. This results in decision models with unreliable predictability and transferability, yet these models are needed to inform policy questions. In this study, we synthesize the attributes of truck activity and parking, including attributes of the tour, parking location, industry sector, and commodity type for the purpose of developing a parking location choice model. This is done through the fusion of four data sources to construct a record of observed parking behaviours of trucks, mainly location choices and corresponding attributes. GPS data are used to construct truck trip diaries across the City of Toronto, including stop location, duration, frequency, and trip time of day and length. Business establishment data are used to infer the industry sectors associated with each tour. Data from the Ontario Commercial Vehicle Survey are used to predict the commodity types associated with each tour. Lastly, parking characteristics are inferred from spatial data representing the features of parking infrastructure.
