This dissertation was written as a part of the MSc in Data Science at the International
Hellenic University. In an ever-changing era, where needs and demands are unstable,
it is important to constantly prevent and solve problems. One of these is public
transport, where the optimization of the timetable is aimed at improving services to
the travelling public. The scope of the study is to optimize the timetable of the New
York City subway by predicting ridership at its stations. By pre-processing the data,
and using different modelling methods, it is possible to investigate in depth the
mobility patterns of the ridership. It is also shown by comparing the results of the
models that the fastest and most effective prediction method in the present study is
the Random Forest Regressor model compared to the others (SARIMAX, Holt
Winters Exponential Smoothing and LSTM).
Collections
Show Collections
