QQESPM-GS: an efficient solution for geo-textual search parameterized by graph-based spatial patterns

Efficient retrieval of geo-textual data is crucial in various applications. While existing research focuses on basic geo-textual queries, such as range and kNN queries, optimizing efficiency for complex queries remains a challenge. This study explores Quantitative and Qualitative Spatial Pattern Matching (QQ-SPM) queries, and showcase their applicability in Point of Interest (POI) search scenarios. This query format allows the search for groups of closely-located and relevant geo-textual objects based on a spatial pattern graph with keywords, pairwise distance limits, and topological constraints. We propose the Quantitative and Qualitative Efficient Spatial Pattern Matching Generalized Solution (QQESPM-GS), an approach designed to efficiently process such queries across different geospatial technologies. QQESPM-GS includes three primary solutions, each one designed to work with specific environments and spatial search engines: QQESPM-Quadtree, utilizing the IL-Quadtree index; QQESPM-EO, applying spatial operations with a greedy strategy; and QQESPM-SQL, enabling efficient execution within a relational spatial database. Experiments with a real POI dataset from UK, show that our approach remains robust with increasing dataset sizes and query constraints. Additionally, it achieves at least one order of magnitude faster query times compared to baseline methods for solving QQ-SPM queries.