Nowadays bike-sharing systems generate large amount of data, what can be used for prediction, estimation, hence creating value for these systems’ operators supporting the distribution and transportation of the bicycles between the stations.
The goal of the thesis is to predict the demand of a docking station at a given time with data mining techniques based on bike-sharing data. This thesis uses data from a challenge organized by the "Big Data - Momentum" research group of the Hungarian Academy of Sciences (MTA SZTAKI) and the Centre for Budapest Transport (BKK), in which the participants analyze data of the MOL Bubi public bike-sharing system’s data.
The second chapter of the thesis is about the project announcement and its detailed interpretation, it gives information where the points are discussed in this document.
It introduces theoretical knowledge about data mining, the most popular methodology used by these projects, and gives information about time series analysis, furthermore about similar publications, and techniques used there.
Then follows the statement of the actual task’s implementation, the dissertation provides information about data, and modelling processes, their efficiency, and outlines some further development, improvement possibilities.
The document concludes by providing information about the new programming techniques, procedures learned during the implementation.