This project was developed as part of the Rupin Academic Center's Hackathon 2023, held on May 11-12. The objective of the hackathon was to analyze taxi traffic data in New York City (NYC) and create a predictive model to estimate the number of taxis at a given time and day.
Our team diligently explored and analyzed the provided dataset, applying machine learning techniques to develop an accurate predictive model. We specifically utilized the K-Nearest Neighbors algorithm, taking into account various factors such as the pickup hour, day, month, tourist season, holiday, season, and rush hour rides.
The model performed impressively on both training and test datasets, achieving a training score of 0.964 and a test score of 0.926. However, the most remarkable result came from the external test, where the model achieved an exceptional R-squared score of 0.941.
To assess the model's performance, we have prepared a CSV file that includes the actual count of taxis, the predicted count, and a score indicating the difference between the actual and predicted counts for each data point. You can find the file here.
We are excited to present the results of our hard work during the hackathon. The predictive model we developed demonstrates its efficacy in estimating the number of taxis in NYC based on various factors. The results obtained indicate the potential value of this model in optimizing resource allocation and enhancing operational efficiency.
For more details about our approach and methodology, please refer to the documentation provided. Feel free to explore the attached CSV file and reach out to us if you have any questions or would like to discuss our project further.
Thank you for your attention, and we hope you find our work interesting and valuable.