High Performance Computing for Big Spatiotemporal Data Mining

Type: Paper
Sponsor Groups: Cyberinfrastructure Specialty Group, Geographic Information Science and Systems Specialty Group, Remote Sensing Specialty Group
Poster #:
Day: 4/12/2018
Start / End Time: 1:20 PM / 3:00 PM
Room: Bayside B, Sheraton, 4th Floor
Organizers: Fei Hu, Jingchao Yang, Gautam Thakur
Chairs: Fei Hu


Earth observation systems and model simulations are generating massive volumes of disparate, dynamic, and geographically distributed spatiotemporal data with increasingly finer spatiotemporal resolutions. Meanwhile, the propagation of smart devices and social media also provide extensive information about daily life activities. Efficiently mining such big data enables us to develop new decision support applications, and provides unprecedented values for business, sciences, and engineering. However, the “Vs” of big data (volume, variety, velocity, etc.) pose significant challenges for discovering the underlying values due to the limitations of traditional data management, computing, and data mining methods. Advanced computing technologies and data mining methods have been closely integrated together and widely adopted to address these challenges. This section aims to capture the efforts on utilizing, adapting, and developing new computing approaches and data mining methods to accelerate geospatial big data analytics in solving environmental and social science problems.

Topics include, but are not limited to:
1. Advanced computing cyberinfrastructure (e.g. GPU, Multicore computing, HPC and cloud computing), frameworks (e.g. leveraging MapReduce, Spark, MPI), and advanced computational libraries/technologies for big spatiotemporal data management and processing.
2. Advanced data mining methods (e.g. geospatial statistics, machine learning, deep learning) for big spatiotemporal data mining, including raster data (e.g. satellite images), vector data (e.g. trajectories), and social media data.
3. Data visualization methods, frameworks, or libraries that help big spatiotemporal data mining.

To present in this session, please send your presenter identification number (PIN), paper title, and abstract to fhu@gmu.edu or jyang43@gmu.edu.

Fei Hu, George Mason University

Jingchao Yang, George Mason University


Type Details Minutes Start Time
Presenter Jing Li*, University of Denver, Mario Lopez, Department of Computer Science, University of Denver, Laura Atkinson, Department of Geography and the Environment, University of Denver, Michael Finn, Center of Excellence for Geospatial Information Science, US Geological Survey, A CUDA-Based Light-Weight Parallel Library for LiDAR Data Processing 20 1:20 PM
Presenter Gautam Thakur*, Oak Ridge National Laboratory, Marie Urban, Oak Ridge National Laboratory, Kevin Sparks, Oak Ridge National Laboratory, Kelly Sims, Oak Ridge National Laboratory, Robert Stewart, Oak Ridge National Laboratory, New Frontiers in Exploiting Open-Source for Geo-spatial Research at Scale 20 1:40 PM
Presenter Xuan Shi*, University of Arkansas, Accelerating Agent Based Modeling for the Simulation of Information Diffusion Using Graphics Processing Unit and Intel’s Xeon Phis 20 2:00 PM
Presenter Andrew Reith*, Oak Ridge National Laboratory, Analysis of Satellite Imagery Characteristics for Convolutional Neural Network Training 20 2:20 PM
Presenter Fangcao Xu*, Penn State University, Spatiotemporal Information Diffusion and Network Analysis for News Media Data 20 2:40 PM

To access contact information login