Improving an Existing Natural Language Bus Route Information System: Adding Support for Geographic Expansion and Visualization in Maps

Abstract

Real-time itinerary information is increasingly used and the functionality of travel planners is becoming more and more advanced. This gives the users the opportunity to tailor their demands and receive simple and fast answers, either through a mobile app, website or voice mail. This thesis explores how the BusTUC system can be improved to take the needs of the users into account. To investigate this, the system is expanded to return answers also for tram departures, and support for visualizing routes in a map is added. 14 persons have tested the new system on a website. Then, the same persons have given feedback through a survey. These answers have been analyzed to discover problems in the web client, and to detect how the experience on the website is, in terms of response time. It seems that correspondence between different means of transportation is important to the users, but this functionality is not present in the current system. Several clients want the system to take all of Norway into account, not just Trondheim. The support for tram schedules and the map functionality seem to increase the credibility of the BusTUC answers, which again lead to more satisfied users. The report also describes how BusTUC can be used in another domain to return intelligent answers. This is an aspect that should be further explored in the future.