Estimation of origin-destination matrix from smart card data of public transportation of the Community of Madrid

Resumen

The increase in the population of the cities brings the growth in the public transportation demand with itself which requires achieving the most efficient and faultless public transportation management. One way to accomplish this is to forecast the citizen mobility flows and understand the traveling behaviors of the passengers. In this thesis, origin-destination (OD) matrices are introduced to fulfill these objectives for the Madrid metropolitan area. Origin-destination matrices are constructed by aggregating the origin and the destination of the trips. To generate the OD matrix for the public transportation networks with entry-only automated fare collection systems such as the the one in Madrid metropolitan area implies the need for destination estimation of each trip. In this thesis, an algorithm is developed by utilizing the trip chaining method to detect the transfers and predict the destination location of the trips performed by the senior citizens of the Community of Madrid. A sample of data belonging to 3 different days is used to prove the algorithm and destinations of around 29% of the trip are estimated. Moreover, the analysis extracted from the obtained OD matrix is demonstrated. From the findings of the thesis and the trip chaining algorithm created dedicatedly for the Madrid metropolitan area public transportation network, insights for the future development of the public transportation network and additional knowledge such as train loads, crowd analysis, and travel forecast can be acquired.