This research focuses on predicting the hourly number of bikes needed using Citi bike data. Micro mobility is the new trend that serves the transportation sector in any city. With the development of technology and introduction of new modes, comes new challenges. Bike sharing is the most developed and standard micro mobility device with extensive data sources. In this research we introduce the rebalancing bike sharing problem, which is very recent and interesting problem. Bikes are being ridden from a station and returned to another, not necessarily the same one of departure, this procedure can cause some stations to be empty while others to be full, as a result, there is a need for a method by which distribution of bikes among stations are done. Using year-round historical trip data obtained from one of the famous bike operators in New York that is Citi bike. The study aims to find the factors affecting bike ridership and then by utilizing some predictive algorithms such as, regression models, k-means, decision trees and random forest a model will be created to estimate the number of bikes needed in an hourly basis regardless of any specific stations initially. Where accuracy will be eventually calculated. The testing will be initially evaluating the data of Citi bike in New York, however, the same can be utilized to evaluate data from other cities worldwide and operators, as well as other micro mobility modes such as e-scooters, mopeds, and others. Initially the Prediction problem will be evaluated against the current data available in the open-source Citi-Bike data, however, weather factors, bike infrastructure, and some other open-source data can be integrated for better results.
Professional Studies (MS)
Department, Program, or Center
Graduate Programs & Research (Dubai)
Asa'ad, Eman, "Hourly Demand Prediction of Shared Mobility Ridership" (2022). Thesis. Rochester Institute of Technology. Accessed from