E-commerce sites, auction sites, financial institutions, insurance companies and telephone companies all have event based data that describes transactions between customers (Social Networks) that are located in time and space (GeoTemporal).

All these transactions together form interesting social graphs and patterns of customer behaviour. Some of these behaviours are very interesting from a marketing perspective, other behaviours might point to fraudulent actions. Analyzing graphs and geospatial oriented data is notoriously hard to do with typical big data solutions, such as Hadoop, so we use a hyper scalable graph database to perform this analysis.

We will present a number of new technologies to make it very straightforward and user friendly to analyze behavioral patterns. We discuss extending SPARQL 1.1 with a large number of magic predicates for geospatial, temporal and social network analysis so that non-specialists can very easily build very powerful queries. We will present new visual discovery capabilities to GRUFF, a graphical user interface for Graph Search. We will demonstrate how users can explore visual graphs and easily turn interesting patterns into SPARQL queries.