Mobility Analytics
Data Modeling
As a Data Scientist hired by Sigma Cab Private Limited, the objective was to construct a predictive model aimed at anticipating surge pricing types.
The company, operating as a cab aggregator service, sought to enhance efficiency by matching appropriate cabs swiftly with the right customers. The need for a proactive predictive model stemmed from the desire to optimize cab allocation and ensure customer satisfaction.
Observation & Findings:
Model Development Approach: The LGBMClassifier was employed as the predictive model for surge pricing type anticipation. The model showcased promising results with an average accuracy of approximately 70.67% across multiple iterations, demonstrating its capability to predict surge pricing types.
Model Performance Metrics: The model's performance, gauged through accuracy measurements, hovered around 70.67%, indicating a reasonable predictive capability. However, there were variations in performance across different folds, with the highest accuracy reaching approximately 71.25% in one instance.
Challenges and Areas for Improvement: Despite achieving a moderate accuracy rate, there remains room for model refinement. Addressing the variability in model performance across folds and enhancing accuracy would significantly improve the model's predictive power.
Application Potential: Implementing this predictive model within Sigma Cab's operations could lead to more efficient cab allocations, optimizing resources, and improving customer satisfaction. However, fine-tuning the model to achieve a more consistent and higher accuracy rate is imperative for its effective integration into the company's operations.