Analysis and prediction of bike sharing usage in New York City

Starting Point

• CitiBike is a large bike sharing provider in New York City
• With over 7,000 bicycles and more than 328 stations in the urban area, around eight million trips were made in 2015
• The team had a wealth of CitiBike data sources to predict future use

Objectives

• Predict user numbers for specific stations to plan marketing measures
• Predict the demand for bikes at specific stations
• Identify areas with a lack of rental stations
• Examine how the proximity of bike stations to subway stations affects the user numbers

Course of Action

• Preparation and unification of existing data
• Execution of a time series analysis in Tableau
• Visualization and clustering of the rental locations
• Calculation of the distance between rental stations and metro stations based on geospatial data

Results and benefits

• Successful proof of concept: visualization of the data and prediction of the use of the CitiBike service
• First ideas for expanding the bicycle network based on actual usage data
• Possibility of forecasting user behaviors for specific bike stations