Social Sensing in Developing Regions: Challenges for Bus Arrival Time Prediction

Abstract

The design of crowdsourcing applications to supplement public transportation information systems have generally assumed availability of high-speed Internet connection coupled with high data sampling and gathering via data-hungry application interfaces. But, in developing regions, low-income users generally avoid the use of data-intensive applications over the Internet connection provided by their mobile operator. Such restriction imposes key constraints and challenges on the design of social sensing applications targetted at low-income communities. In particular, the design of crowdsourcing applications for bus arrival time prediction in developing regions should seek high accurate prediction based on minimal data gathering and infrequent data sampling.