The sheer scale of very large data sets has resulted in algorithmic techniques that sample small ‘slices’ of the data set, in order to begin the task of understanding patterns and anomalies. In effect this is a form of abductive reasoning, where a starting proposition is tested, which in turn informs subsequent iterative sampling until some critical insight is reached. Parallel to research in Big Data the architectural discipline has developed a form of abductive reasoning undertaken with visual representations (2D and 3D models). Given the recent interest within Big Data discourse for the use of augmented and virtual reality to undertake visual analytics, we see potential in adapting the abductive spatial reasoning from architectural design. How might we conceive a user interface for exploring big data in AR and VR, which allows intuitive spatial sampling that leverages the spatial acuity of architects and associated disciplines?