2023-08-30, 09:00–10:30, Room 18 (Sala 18)
High-resolution, continental-scale modeling enabled by modern, massive datasets, requires development of scalable geoprocessing workflows. To enable participants to effectively use available computational resources (laptop, desktop, institutional HPC), we will introduce basic parallelization concepts such as parallelization efficiency and scaling. We will explain various approaches to parallelization in GRASS GIS, an open source geoprocessing engine, that rely on OpenMP, Python and Bash.
In the hands-on part, participants will speed up an urban growth model by parallelizing different parts of this complex geoprocessing workflow using techniques that are easily applicable to a wide range of analyses and computational resources. The workshop will be running in a Jupyter Notebook environment using GRASS GIS Python API to run GRASS tools and visualize results of the analysis in a reproducible way.
Participants will be able to either run the workshop on their laptops (see instructions) or in a cloud environment (using WholeTale, no installation required).
Caitlin is a 3rd year doctoral student in the GeoForAll Lab at North Carolina State University in Raleigh, NC, USA. She has been working on improving the integration of GRASS GIS and Jupyter Notebooks.
Anna is a geospatial research software engineer with PhD in Geospatial Analytics. She develops spatio-temporal models of urbanization and pest spread across landscape. As a member of the OSGeo Foundation and the GRASS GIS Project Steering Committee, Anna advocates the use of open source software in research and education.