Efficient Support of Text and Time in Spatial Data Systems

Ahmed R Mahmood, Purdue University


The widespread use of GPS-enabled devices, e.g., smart-phones and GPS navigators has resulted in the generation of massive amounts of spatial data. This has led to the development of spatial data systems that process and manage spatial data efficiently. Temporal and textual data often coexist with spatial data. This dissertation targets the efficient processing of spatio-temporal and spatio-textual data. To efficiently support spatio-temporal data with limited temporal data retention, a new disk-based data structure for indexing moving objects trajectories with limited temporal data retention is introduced. To optimize the processing of spatio-textual data, a new frequency-aware spatio-textual indexing structure is presented. The proposed spatio-textual index adapts to the differences in frequencies of the indexed keywords over space and time. To improve the scalability of spatio-textual processing, a new distributed spatio-textual stream processing system is presented. The proposed system fairly distributes the workload and co-locates the data objects with the corresponding queries at the same worker processes. By applying dynamically evaluated cost formulae that continuously represent the processing overhead at each worker process, the proposed system adapts to changes in the workload and ensures fair workload distribution across worker processes. Finally, to formalize querying spatio-textual data, a new query language is introduced to express complex spatial-keyword queries. The proposed query language uses declarative spatial and textual building-block operators and predicates to represent a wide range of spatio-textual queries.




Aref, Purdue University.

Subject Area

Computer science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server