Dynamic Update of Public Transport Schedules in Cities Lacking Traffic Instrumentation

Document Type


Publication Date



A common obstacle for citizens in switching to public transportation is the lack of information about available choices when they need to travel. Al-though schedules of individual modes like bus or metro may be available as paper pamphlets, or digital files on websites, they do not give an integrated view of the complete services possible when a citizen actually wants to travel. Furthermore, the situation on the roads evolve and this demands timeliness of public transportation information. We want to tackle this problem in the context of cities of developing countries like India, which lacks basic instrumentation to track road conditions or vehicle location. Our solution extends a public transportation recommender working only with static schedule information to utilize SMS messages about road conditions sent by city authorities. Our solution consists of: (a) extracting events from traffic alert messages, (b) reasoning about traffic delays from extracted events by qualitatively deciding what stops (locations) will be affected, and quantitatively estimating the lower bound on the probability of having a delay at those locations in the city, and (c) utilizing the delay estimates for route recommendation. We use publicly available traffic related SMS messages from Delhi, India, to evaluate our approach and show its promise. Our solution provides dynamic updates for transport network in cities with low investment and quick time to realization.


IBM Research Technical Report RI14002