Please note that the demo apps are not meant to be end user apps so they may be buggy and not user friendly. The main purpose of these apps are to show basic uses of the project for others to incorporate in their better prepared transit apps.
The Predict Demo App:
This app demos predicting how long a bus will take to get to a stop. To use it select a bus then select a stop. It only shows route 42.
http://hestia.nighsoft.net/septa_next_bus/demo/snb_demo.html
The Next Bus Demo App:
This app lets you select a stop and then it will return the next buses for each route which will be at that stop and the predicted times for them to get there.
http://hestia.nighsoft.net/septa_next_bus/demo/snb_demo_next_bus.html
The Nearby To Demo App:
This app lets you select a from location and a to location then it will determine all the routes which go in between those stops and return the buses in order for each route that will take you from the initial location to the destination. It will also provide the predictions for each bus to get to each stop.
http://hestia.nighsoft.net/septa_next_bus/demo/snb_demo_nearby_to.html
Training Performance Results:
This app allows you to see graphs and results of the machine learning algorithm for the routes and stops. These results are used to find anomalies and bugs in the training and to see which routes / stops were able to train good etc. A route / stop with a low "mean squared error" normally means that it will make good predictions.
http://hestia.nighsoft.net/septa_next_bus/demo/snb_performance.html