A taxi client aims to develop a machine learning model that can accurately predict whether a trip will be rewarded with high tip. Additionally, the client is also interested in understanding the key factors influencing tip amounts. The insights gained from significant features will be utilized to inform and recommend strategies for enhancing taxi operations.
The data was taken from Kaggle New York City Taxi Trip and New York City Taxi Fare Prediction
| Conclusion | Suggestion |
|---|---|
| Target at the right time to meet customer’s demand | Define ‘peak-day’ and ‘peak-time’ to allocate more resources to meet the demands: • More call center reps to take trip requests • More incentives for drivers during peak times • Rate adjustment during peak times Spread out the peak demand to other time frames by applying promotion codes for the duration before and after peak times. |
| Performance is similar between vendors | To increase competitiveness, consider partnering with a third vendor. Conduct quarterly/monthly performance reviews for vendors for improvement. |
| Passenger demand reduced significantly in harsh weather, but not the tip | Ensure safety while encouraging more drivers to take trips during harsh weather by offering incentives. |
| The taxi operation is popular in certain neighborhoods. | For other areas, there is still room to serve more passengers and gain market share. Allocate more drivers to those areas with moderate demand and high tips (>20%). Promote presence in those valuable neighborhoods. |
- Collaborate with Marketings to plan the suitable approach for the potential customer.
- Present these insights to Partners Service Team and get their feedback on how to allocate number of Drivers efficiently.
- Further analysis about customer sentiment on those high/ low tip trip can be conducted with additional information about customer's rating and comment.