Final Project: Predicting Trip Times & Final Destinations of Taxi Trips in Porto, Portugal

General Assembly Part-Time Data Science Course

Completed November 5, 2016

Overview

This project is based off of two Kaggle competitions (I and II) from 2015 that were posted by ECML PKDD* and completed by conference attendees.

The competition was a demonstration on how to make electronic dispatching systems more efficient by predicting the earliest time in which a driver can accept a new passenger based upon the location of the final destination and a prediction on how long the trip should take.

The Kaggle/ conference competitions had two immediate aims:

1. Predict the final destination of a taxi ride in Porto, Portugal based upon a partial trajectory of the beginning of the ride.
2. Predict the duration of the ride.

*ECML PKDD = European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases

Final Project Objectives

For my final project, I chose to predict the final destination and trip time of a randomly selected partial trajectory (trip #40).

Files

• taxiTripFunctions.py: functions used in the analysis, feature engineering
• FinalProject_ExploratoryAnalysis_Model.ipynb: IPython notebook of complete project