Improving Micro-Mobility: A Case Study of Citi Bike Usage in NYC

This report presents a data science case study using Citi Bike trip data to analyze usage patterns, forecast demand, and support operational planning for a micro-mobility system in New York City. The project addresses the challenge of balancing bike availability with rider demand across time and location. Publicly available data from 2024 and 2025 was processed and analyzed using Python in Google Colab, with a focus on temporal and spatial behavior. 

Exploratory data analysis revealed strong commuting trends, and K-Means clustering identified five meaningful geographic regions of concentrated ridership. Time series forecasting using both SARIMA and Meta’s Prophet models highlighted Prophet as the superior approach in predictive accuracy, although limitations such as overfitting and irregular seasonal patterns were noted. The combined use of forecasting and clustering enabled a dual perspective on system performance by anticipating future demand while identifying current usage hotspots. 

These methods offer practical strategies to improve rebalancing, infrastructure investment, and customer experience.