Scalable Collective Spatial Keyword Query

Spatial keyword queries have been widely studied recently, along with the emergence of large amount of geotextual data. We consider the problem of scalable collective spatial keyword queries in this paper. Such query has a wide spectrum of applications; for instance, to find the best (nearest) area to organize a friend get-together where bars, restaurants and accommodations are nearby, and compose a group of members from different professional domains, e.g., computing, accounting, etc, for a specific task, etc. While existing algorithms processes the queries well, we observe their shortcomings in handling large-scale datasets. To this end, we propose a distributed solution following Spark programming paradigm. Moreover, a grid-based optimization technique is further proposed to enhance the efficiency. Extensive experiments on various datasets confirm that the proposed algorithm efficiently solves the problem at scale.