Tools and Analytical Techniques
This project was created with the following tools:
Python
Google Colab
Streamlit app
Objective:
The analysis aims to optimize Citi Bike's bike-sharing logistics across New York City by examining user behaviour to inform strategic decisions. This project identifies key metrics to improve bike availability at high-demand stations and streamline distribution, reinforcing Citi Bike’s commitment to sustainable urban mobility.
“Interactive Citibike Dashboard: Usage Metrics and Forecasting”
A detailed analysis of bike-sharing data, highlighting peak usage, station performance, and user behaviour.
Motivation:
With rising demand post-COVID, Citi Bike faces challenges in bike distribution, including shortages and overcrowding at docking stations. This analysis addresses these issues by uncovering patterns in peak times, seasonal demand, and usage hotspots. It provides management actionable insights to improve customer satisfaction and support future expansion.
Project Scope and Planning
Align requirements, project scope, and desired outcomes of the project.
In this project, I analyzed the New York Citi Bike data to understand customer behavior, rental patterns, geographic distribution, and revenue. I aggregated bike trip data across New York to identify the most popular starting locations and to summarize the data yearly to reveal seasonal patterns. Additionally, I applied geographic plotting techniques to identify problem areas in station distribution and to explore the most common routes taken by riders.
Specifically, I aimed to uncover:
Scaling Back Bikes (November-April)
Ensuring Bike Availability at Popular Stations
Expansion Along Waterfront Areas
By conducting this analysis, I can present insights essential for addressing the distribution issues.
Data Preparation and Exploration
Applied Techniques:
Phyton
Data Import and Preparation:
Data Merging and Integration:
Data Wrangling
Aggregating Data
Exploratory Data Analysis (EDA):
Time Series Analysis
Geospatial Analysis:
Interactive Visualization and dashboard Deployment
Final Insights and Recommendations:
How much do you recommend reducing the number of bikes during the months of November through April?
Analysis indicates that bike usage declines significantly during the colder months, mainly due to weather conditions and decreased tourist activity. Historical data shows that reducing the bike fleet by 20-30% during this period can help lower operational costs while still providing enough bikes for regular users. Additionally, a correlation model between weather and bike usage can aid in determining the optimal fleet sizes for each winter month.
Ensuring the Availability of Bikes at Popular Stations
A key operational challenge is maintaining bike availability at busy stations with high arrival and departure rates. Fluctuations in bike supply are common due to constant movement. Visualizations can identify peak stations and seasonal trends for better focus. Predictive models using historical data can forecast busy times, allowing for proactive bike rebalancing. Real-time monitoring, dynamic teams, and user incentives also help keep these stations well-stocked year-round.
Aggregated Bike Trips in NYC 2022
Waterfronts experience a significant seasonal demand, making them ideal locations for additional bike stations. A spatial analysis of trip data reveals several underserved areas along the waterfront where bike traffic is high but station coverage is lacking. We can enhance service and improve coverage by adding new stations within 500 to 800 meters of these high-traffic locations and implementing pilot programs to assess demand.
Classic vs Electrical Bike Rental by Temperature
The analysis shows that bike rentals rise as temperatures increase. Classic bikes are generally more popular, while electric bikes demonstrate steadier usage in colder weather. Peak demand occurs at temperatures above 15°C, driven by favorable cycling conditions. Electric bikes offer more convenience during challenging weather, allowing for consistent usage even in colder periods. These findings highlight the importance of aligning bike availability with seasonal weather trends to optimize service and effectively meet user demand.
Conclusions and Recommendations :
Conclusions:
The analysis shows that bike rentals increase with rising temperatures, indicating that warmer weather attracts more users. This pattern emphasizes the importance of adjusting operational planning to meet seasonal demand.
Classic bikes typically have higher rental frequencies, and electric bikes exhibit more consistent usage, particularly during colder weather. This suggests that users prefer electric bikes in less favourable conditions, likely due to their greater ease of use.
The most frequently used start and end stations reveal consistent high traffic at certain locations. This insight helps optimize bike placement and inventory management, ensuring popular stations are adequately stocked to prevent shortages or oversupply.
Recommendations:
Our analysis indicates that NY CitiBikes should focus on the following objectives moving forward:
There is a clear correlation between temperature and bike trips. I recommend ensuring that all stations are fully stocked during the warmer months to meet the higher demand while decreasing the number of bikes in circulation during winter and late autumn to reduce costs.
Reducing the fleet by 20-30% during the November to April period can help minimize operational costs while ensuring there are enough bikes available for regular users.
Stations along the waterfront and around Central Park show a higher popularity. I recommend increasing the number of bikes and bike parking options at these locations.
Classic bikes are rented more than 2.5 times as often as electric bikes, primarily due to the limited availability of electric bikes. I recommend incorporating more electric bikes into circulation when adding new bikes.
Dataset
Source:
Weather data using NOAA’s API service.: NOAA
Open source data from the Citi Bike database for 2022.: Citi Bike Trip Data 2022