Data analytics based model to estimate ride-sharing potential in Sri Lanka

Abstract

Traffic handling in cities is becoming a major issue worldwide. Everyone is in a hurry to go to work or deliver goods on time. The more the vehicular traffic, the higher the pollution. This is a waste of time, fuel and energy. Currently, there are many types of research being done to come up with a solution to reduce vehicular traffic. Ride-sharing is one of the potential solutions to reduce the traffic congestion by reducing the number of vehicles entering a city. The idea of Ride-Sharing is to share rides to/from home/work daily, based on home/work locations. Identification of home/work locations is one of the major task in Ride-Sharing to identify potential ride-sharers. Identification of these locations can be done using CDR data. There are models and algorithms that have been introduced by several types of research to identify the home/work locations based on CDR data. However, this has not yet been implemented in Sri Lanka. The idea of this study is to identify the potential of Ride-Sharing in Sri Lanka using CDR data. End-Point and En-Route Ride-Sharing are considered as the main Ride- Sharing options. Analysis was performed on data collected in 2012/2013 period for 41 cities of the Western Province of Sri Lanka. To identify the home/work locations, the hours between 21.00-05.30 was considered as home hour events and the hours between 10.00-15.00 considered as work hour events. As per the analysis based on the collected data it was identified that there are 72.94% potential ride-sharers and based on the transportation data it was identified that there are 38.43% Private transportation modes users in the selected cities/towns. Hence, it was identified, that there is a potential of implementing Ride-Sharing and it has a high impact on traffic congestion. The decision was mainly based on the number of vehicles entering the cities.