Generating Regular-Interval Maps from Public Transit Data
Type: You should start working on this topic as a bachelor or master project to test the waters. If you find you like the work and want to improve your results after the end of the project, you can of course continue to work on it as a thesis. You should be fond of big datasets, timetable data and graphs.
Background info: We are developing an approach to draw transit maps automatically (demo). At the moment, the transitmapper uses GTFS data to render the lines that travel between two stations (just as a normal transit map on some station displays it). Another possibility (useful for larger train networks) would be to display all the routes that travel through the network together with their regular intervals, as they occur for example in train networks serving an "Integrated Synchronised Timetables (Integraler Taktfahrplan)". In the map, instead of displaying a line between two stations with its name, the line is now displayed with its regular interval (every 10 minutes or every 20 minutes an so on). If two or more intervals are interlaced, display both of them separately. See this example of the entire Swiss train timetable, displayed as a map.
Goal: Your task would be to take some timetable data in the form of GTFS feed, generate a graph which contains each trip occurance between two station and analyze these occurances for frequency covers. (For example, if a station is served at minute 14, 29, 44 and 59, the frequency cover is "15 minutes"). See [1] for an efficient algorithm to do this. You should make sure that if a node (a station) has multiple edges, some frequency occuring on some edge can be assigned to its continuation in another line. After that, you should do some sensible grouping of lines and output the graph as a GeoJSON-File which will be specified by us and which can be used an input for the transitmapper. You can use a programming language and toolsets of your choice.
Optional goal: Provide some simple web application where your (unrendered, raw!) dataset can be looked at on a map.
[1] H. Bast, S. Storandt. Frequency-Based Search for Public Transit. SIGSPATIAL 2014