ST-optics: a spatial-temporal clustering algorithm with time recommendations for taxi services

Abstract

Taxis have become the most popular and preferred mode of transport because of the comfort and convenience they offer. In this work, we design a novel algorithm, called ST-OPTICS, which helps taxi fleet owners to manage their cars by recommending the nearest hotspot of pickup locations based on the current time. It will be useful for personal as well as shared taxi drivers. ST-OPTICS is a density-based algorithm to find spatio-temporal clusters by analyzing the pickup trends of spatially close locations. In contrast to the existing clustering algorithm, it handle noise points in a different way and forms clusters of varying density based on spatial and temporal closeness. Recommendations from ST-OPTICS will consist of locations along with the timings during which they’ll behave as pickup hotspots. This algorithm was evaluated on Singapore’s taxi cab data. Our results indicated a superior performance of the proposed ST-OPTICS algorithm.