The client is a subsidiary of Roads and Transport which provides its customers with a range of services, comforts, and customer care, with the highest levels of transportation quality standards.
- High percentage of non-revenue trips: Taxi spends half of its operational time on non-revenue trips, and the other half have riding customers.
- Commercial viability: Operational cost against revenue is high, which in the long run may limit business development
- Technology and innovation: Lack of taxi demand visibility barricades adequate provisioning at the ‘right place’ and at the ‘right time’.
- Limited knowledge of the taxi demand: Limited data and visibility resulted in inefficient operations with long passenger wait time and several vacant trips
A state-of-the-art AI/ML-based predictive analysis system for forecasting the Taxi Demand at any given future time, based on various inputs along with integration with various systems/sources running in the client environment.
- A web application that provides future predictive analysis statistics and various other useful information to help critical business decisions
- Android mobile application to accept requests and earn incentives on the trips drivers make in high-demand areas
- Prediction of taxi demand from historical and real-time data sources
- Prompt the most optimized route for the driver through a mobile app with the highest probability of getting rides
- Dynamically re-optimize & propose new route, when demand drops in the area of taxi movement
- Help the operations team to make micro-decisions based on the predicted demand in future
Tools & Technologies
- Power BI
- Cosmos DB
- Azure Data Factory
- Azure ML, H2O.ai
- Azure IASS/ PASS
- Improved taxi engagement ratio (Engaged/Vacant)
- Reduced passenger wait time
- Reduced idle time of the driver with quick updates
- Augmented revenue/non-revenue kilometer ratio
- Improvised revenue/kilometer
- Increased number of revenue trips
- Advanced analytics reports improved decision making