Improving spatial search using a graph database and Uber H3 / Jaye Renzo Montejo, Irvin Kean Paulus Paderes, Ligayah Leah Figueroa.

With the rise in popularity of applications about ride-sharing, food delivery, courier service, and any subject that caters in fulfilling the needs of an on-demand economy, access to spatial information in real-time becomes an important aspect. To achieve this, spatial databases are used to store and process location information that are consumed by these applications. Normally, these databases use relational modelling. This research proposes the use of hexagonal discrete global grid systems in graph databases in order to improve upon existing spatial databases. In measuring the query performance, a geofence surrounding Quezon City, Philippines was created and various amenities within the area were extracted then used as the dataset. Bounding box search queries were then performed on this dataset. A working implementation of a graph-powered spatial database with a hexagon discrete global grid system called H3 was presented, and it was shown that this setup can pro-duce lower query execution times at scale than those of relational databases