CSCI 4502 Data Mining Project
Cristian Mendivil
Luke Lyon
Theodore Freeman
Vamshi Arugonda
Ride-sharing on Uber and Lyft has exploded in popularity in the past decade. The growth of these companies has far surpassed legislators' abilities to regulate them and city planners' abilities to design cities around the services. Public transit has been a staple of Boston since 1947, with the creation of the Massachusetts Bay Transit Authority (MBTA) \cite{wikipedia_2018}. We aim to use Uber travel times in Boston as well as data provided by the MBTA on the reliability of rail, subway, and bus routes to establish a link between delays in MBTA service and an increase in traffic times. With this link, we then hope to train an algorithm to predict delays in traffic given delays in public transit.
- Do delays in public transit times correlate with increased travel times for uber?
Yes, it is a positive correlation for bus routes and uber rides. - What are the most influential weather patterns affecting bus travel time?
We don't know. Our regression algorithms failed to create an accurate predictor for percentage of buses on time given weather data. - Based on the delay from weather patterns, how much extra travel time should a Boston commuter account for?
We were unable to answer this as well due to the regression algorithm failure.
Uber drivers could use the MTBA real-time data streams to try to forecast what their delays might be and plan their routes accordingly
We could look for further data to incorporate into building a predictive model for percentage of buses on time
https://drive.google.com/file/d/1H8wA2UIywPcSQc73qefP-CJBq5NIfxW3/view?usp=sharing
https://github.com/RadiationMiners/bostonubertraffic/blob/master/09_BostonUber_Part4.pdf
(View-Only link to Google Doc) https://docs.google.com/document/d/1J2WKNe6biJlcB8gUCP5kc_z4dGGvhAotPs_rCJMkZKo/edit?usp=sharing