Apparently I’m up and running.
When trip breaking has been performed on the database of collected GPS points, it is useful to know which roads a user has traveled along. Any person could easily figure out the routes taken by overlaying GPS points on a street map. Looking at our point set, it’s easy to tell which direction a user is going if they’re driving on the right side of the highway, but explaining that to a computer is another matter. Luckily, this network type of driving network can be simply encoded using edges and nodes. Each street constitutes an edge and each intersection is coded as a node (point). Each edge contains information about which nodes it is connected to as well as legal vehicle travel direction.
An app for learning about travel patterns using a smartphone does not to be complicated. Its main purpose will be to act in the background recording GPS coordinates and occasionally uploading to a database. Battery life optimizations will be different across various makes and devices, but especially phone operating system. Apple’s iOS is more efficient with a limited pool of hardware to support, but most phones are Android (2015). The specifics of an app are not important for this post, the only concern are the characteristics of the points themselves. Generally this means priorities are collection frequency—resulting in higher resolution tracks, precision—using GPS over cell-tower triangulation or wifi, and a timestamp.