πŸš•

TaxiMize

Taxi Drivers AI Assistant

πŸš• What isΒ TaxiMize?

TaxiMize is an intelligent mobile application designed to help NYC taxi drivers maximize their earnings by minimizing idle time and optimizing ride selection. The app leverages machine learning to provide data-driven insights that transform how drivers approach their work.

πŸ”— Project Resources

My Role:BackendDevOps
πŸ“GitHub Repository
View complete project β†’
βš™οΈMy Backend Work
View my contributions β†’

πŸš€ Maximizing Driver Profits

Numerous studies show that taxi drivers often struggle with low earnings due to inefficient ride selection and excessive idle time. Our primary goal with TaxiMize is to help drivers maximize their profits.

The core formula driving our solution
flowchart TD DP[πŸ’° Maximize Driver Profits] DP --> ME[πŸ“ˆ Maximized Earnings] DP --> MIT[⏱️ Minimized Idle Time] ME --- MEDesc["<small>Maximize the earnings per ride</small>"] MIT --- MITDesc["<small>Minimize the time where taxi<br/>is on shift without a passenger</small>"] style DP fill:#1f2937,stroke:#374151,stroke-width:3px,color:#ffffff style ME fill:#22c55e,stroke:#166534,stroke-width:2px,color:#ffffff style MIT fill:#ef4444,stroke:#991b1b,stroke-width:2px,color:#ffffff style MEDesc fill:transparent,stroke:none,color:#666666 style MITDesc fill:transparent,stroke:none,color:#666666

To achieve this, we built 3 smart features:

πŸ—ΊοΈ
Smart Zone Recommendations

Our machine learning models analyze passenger demand patterns across NYC's 264 zones, predicting passenger availability by hour, day, and month.

ML Algorithm: Analyzes historical pickup data to calculate passenger count predictions for strategic positioning.

⭐
Intelligent Ride Scoring

Before accepting any ride, drivers receive a 1-5 star rating based on our machine learning model that evaluates ride profitability and drop-off zone quality.

Smart Evaluation: Drop-off zone "hotness" prevents dead-heading back to busy zones.

πŸ•
Optimal Shift Timing

The app provides recommendations on when to start working, helping drivers avoid unprofitable periods and capitalize on high-demand times.

Development Note: Currently uses hardcoded hours, planned for future ML implementation.

πŸŽ“ Academic Project Context

TaxiMize was developed as the final project for my Computer Science Master's degree at University College Dublin (UCD). This project was supported by extensive literature review and represents a new approach to taxi driver optimization.

πŸ“š
Research Foundation
  • β€’Extensive literature review supporting the approach
  • β€’Novel solution - no similar comprehensive app existed
  • β€’Academic rigor meets real-world application
  • β€’Written report available soon
πŸ‘₯
Collaborative Development
6-person team over 2 months
2 Frontend developers
2 ML engineers
1 Backend/DevOps (My role) ✨
1 Literature researcher