Integrating Traffic Data Stream with Public Route Planning Algorithm and Personalizing it for the End Users

Route suggestions for public transportation has been integrated to different mapping services such as Google Maps or Bing Maps. However, those suggestions are often found to be unreliable for the end-users, mainly because of two reasons. Firstly, these route-finding techniques do not consider traffic conditions on the roads. Secondly, routes are generalized for every user, completely ignoring their commuting patterns. We propose an architecture with a modified routing algorithm that aims to solve both the problems. Our system incorporates both real-time and historical traffic-data into public-transit routing framework. We also personalize transit routes for users by analyzing their commuting behavior. By experiments, we show that our system significantly improves the quality of route-suggestions compared to stateof-the-art techniques with minimal overhead (as low as 20 ms for a 3 km long route).