Short Course at ICSA 2025
2025 ICSA Applied Statistics Symposium, University of Connecticut, Storrs, CT, USA, June 15-18, 2025
Course Description
This course, designed for data scientists, geospatial analysts, and researchers, provides a comprehensive understanding of advanced methods in large-scale geospatial data science. The focus will be on three key topics: large-scale data modeling and prediction, accelerating geospatial data processing with multi- and mixed-precision techniques on modern hardware architectures, and parallelizing related R codes using the first parallel runtime system package in R.
High-Performance Computing Showcase
Transform Your Geospatial Data Processing: Witness the power of high-performance computing in geospatial simulation!
Snail-Paced Processing
Watch "Oppenheimer" and still have time left ⏳🎬
Lightning-Fast Processing
Blink and you'll miss it ⚡
In this course, you'll learn cutting-edge techniques to dramatically accelerate geospatial data processing, reducing computational time from hours to mere seconds!
Instructors
Course Topics
-
Overview of Spatial Statistics (1 hour)
Introduction to spatial statistics, including background and tools for large-scale spatial data manipulation.
Instructor: Marc Genton -
Introduction to High-Performance Computing (HPC) (1 hour)
Overview of HPC and parallel hardware systems with ExaGeoStatCPP introduction.
Instructor: Sameh Abdulah -
Hands-on: Spatial Data Modeling (1 hour)
Practical session using ExaGeoStatCPP in R.
Instructors: Sameh Abdulah & Mary Salvaña -
Multi-Precision Computing (1 hour)
Overview of MPCR R package for CPU and GPU architectures.
Instructor: Mary Salvaña -
Hands-on: Multi-Precision Computation (1 hour)
Practical session with MPCR package.
Instructors: Mary Salvaña & Sameh Abdulah -
Parallel Processing with RCOMPSs (1 hour)
Introduction to parallel processing in R.
Instructor: Sameh Abdulah -
Hands-on: Task-Based Algorithms (1 hour)
Practical session on parallelizing spatial data analysis.
Instructors: Sameh Abdulah & Mary Salvaña
Course Resources
ExaGeoStatCPP: Overview
The power of modern computing for spatial data science
Course Introduction
Overview of the Large-Scale Spatial Data Science course
Spatial Statistics Overview
Introduction to key concepts in spatial statistics
High-Performance Computing
Overview of HPC concepts and applications
High Performance Computing in Geostatistics
Understanding how HPC can be used to accelerate geostatistics computations
Prerequisites
Background in data science, geospatial analysis, or related research. The course is designed for individuals interested in advancing their skills in large-scale geospatial data science, specifically those interested in geospatial data modeling, multi-precision computing, and parallelization of R code for high-performance computing.
Duration
7.5 hours