BusTripper is a project in development to improve transit scheduling and provide automated real-time arrival predictions. Using GPS data from Android devices onboard a fleet of buses, we can identify their concurrent routes and compare their locations against a published schedule. BusTripper will learn from historical data to optimize routes and wait times, and aid in the synchronization of transit systems to prevent issues like bus bunching.
BusTripper is being developed as part of a data science class project at UC Berkeley in coordination with Via Analytics.
BusTripper currently uses the following packages:
python, and common libaries like numpy, pandas, sklearn, matplotlib, re, and sqlite3
R, the DTW package for R, and the python wrapper rpy2
shapely (optional) is used for a visualization tool.