SpaceTime: SpatioTemporal Data Exploration

There has been a meteoric rise in the use of spatiotemporal data in a variety of fields of science and engineering. This growth has been enabled by two complementary trends. First, the newfound ability to deploy sensors, IoT devices, “smart dust”, and mobile phones, at scale have empowered scientists and practitioners to collect unprecedented amounts of location-based data. Second, the computational ability to process and act on such scales of data has also seen massive improvements in the recent past. The combination of these two trends has led to simultaneous technological revolutions in multiple areas of inquiry, such as precision agriculture, smart city management, fine-grained climate modeling, disaster management, and autonomous transport.

We propose SpaceTime, an interactive spatiotemporal framework is built on top of different research projects and software suites. Spacetime utilizes an efficient and interactive computational back-bone for storing and retrieving spatiotemporal data through our GeoSpark system, an in-memory cluster-computing framework for processing large-scale spatiotemporal data. While GeoSpark guarantees efficient fetching of results, a session-based framework is needed to manage ad-hoc query requests at frontend as well. For this purpose, we employ Dice as a distributed system for session-oriented exploration of spatiotemporal data.


  • Behrooz Omidvar-Tehrani, Arnab Nandi, Seth Young, Nicholas Meyer, Dalton Flanagan: DV8: Interactive Analysis of Aviation Data, ICDE Demo, 2017
  • Sobhan Moosavi, Rajiv Ramnath, Arnab Nandi: Discovery of Driving Patterns by Trajectory Segmentation, ACM SIGSPATIAL PhD Symposium, 2016