- Provided by: Geospatial Research Working Group
- Registration: Register Here
On November 6-7, 2024, the SCINet Geospatial Research Working Group (GRWG) will be hosting their 2024 Annual Workshop. The workshop is split over two half-days that will have sessions including:
- Lightning talks on geospatial analyses and workflows that use SCINet resources
- Hands-on tutorials that cover a variety of topics:
- Applying computational methods such as environmental niche modeling and graph convolutional neural networks
- Using Python packages for multidimensional arrays to process geospatial data
- Migrating your workflow to SCINet
- Discussions on GRWG activities for the new year, including desired training topics and collaboration opportunities.
Workshop Goals
The two overarching goals of the GRWG Annual Workshop are to:
- Provide hands-on tutorials on workflows to access the SCINet high-performance computing (HPC) systems and conduct geospatial research at scale.
- Foster research efforts that had previously been unattainable due to computational limitations or technical bottlenecks.
Organizing Committee
The organizing committee for the 2024 Annual Workshop comprises:
- Andrea Albright, SCINet Postdoctoral Fellow
- Amira Burns, SCINet Fellow
- Amy Hudson, Research Ecologist and working group co-lead
- Erika Peirce, Remote Sensing Specialist
- Lavida Rogers, SCINet Training Coordinator
- Heather Savoy, SCINet Computational Biologist and working group co-lead
- Brian Stucky, SCINet Computational Biologist and acting CSIO
- Melanie Veron, SCINet Fellow
How to Participate
To participate in the workshop, please register by submitting the registration form. A calendar event with the Zoom call-in information will be sent after you register.
The form asks for your SCINet account since full participation in hands-on activites will require an account. If you do not have a SCINet account already, please apply for one. We recommend applying for an account as soon as possible, as the process can take time for final approval. All registrants will be invited to a pre-workshop support session the week before the workshop to test logging into their SCINet account.
Pre-Workshop Instructions:
To help minimize technical issues and delays at the start of the workshop, please try the following four tests prior to the workshop.
- Logging on to Ceres Open OnDemand (OOD): Please confirm you can successfully log in to Ceres OOD with your SCINet account (see instructions here). If you are successful, you will be able to see the Ceres OOD home page.
- Ceres Shell Access: When on Ceres OOD, click on the top navigation bar: “Clusters” > “Ceres Shell Access”. A new tab will appear that looks like a shell terminal (e.g., like PowerShell). Please confirm you do not receive any error messages or requests to re-authenticate and that the final line looks like “[firstname.lastname@ceres ~]$”.
- RStudio Server: Back on the main Ceres OOD tab, click on the top or side navigation bar: “Interactive Apps” > “RStudio Server”.
- Fill the input fields with the following (input fields not listed below can be left at their default values):
- Queue: short
- Number of hours: 1
- Number of cores: 2
- Memory required: 6G
- Optional Slurm Arguments: (leave empty)
- Click the “Launch” button.
- Wait a moment for the job card to update from “Queued” to “Running”.
- Please confirm that clicking on the “Connect to RStudio Server” button opens a new tab with the RStudio Server interface.
- Fill the input fields with the following (input fields not listed below can be left at their default values):
- JupyterLab Server: Back on the main Ceres OOD tab, click on the top or side navigation bar: “Interactive Apps” > “Jupyter”.
- Fill the input fields with the following (input fields not listed below can be left at their default values):
- Queue: short
- Number of hours: 1
- Number of cores: 2
- Memory required: 6G
- Optional Slurm Arguments: (leave empty)
- Jupyter Notebook vs Lab: Lab
- Working Directory: $HOME
- Click the “Launch” button.
- Wait a moment for the job card to update from “Queued” to “Running”.
- Please confirm that clicking on the “Connect to Jupyter” button opens a new tab with the JupyterLab Server interface.
- Fill the input fields with the following (input fields not listed below can be left at their default values):
Workshop Sessions
-
Kickoff with lightning talks
This session will have an introductory presentation about the working group and workshop goals plus lightning talks by ARS researchers and SCINet fellows about their research.
-
Demonstration: Species distribution modeling with MaxEnt
During this session, we will explore the topic of species distribution / environmental niche modeling (SDM/ENM), with a particular focus on MaxEnt (maximum entropy), a popular SDM/ENM algorithm used to estimate species habitats from presence-only occurrence records.
-
Tutorial: Geospatial code development using VS Code and Python
This tutorial will illustrate how to use Visual Studio (VS) Code, a flexible development environment available on SCINet’s Ceres cluster, to develop Python scripts. By the end of the tutorial, participants will know how to access VS Code in Open OnDemand and use it to navigate files on Ceres, access Python packages, and develop and execute Python scripts.
-
Tutorial: Applying geographically informed Graph CNNs in disease ecology modeling
This tutorial presents a workflow using a Graph Convolutional Neural Network with LSTM layers (GLSTM model) to classify West Nile Virus disease presence or absence in horses across counties in a subset of states in the US
-
Tutorial: Using xarray to work with multidimensional geospatial datasets
In this tutorial you will learn the basic concepts of the Python library
xarraywhile exploring its capabilities when working with multidimensional datasets with dimensions of time, latitude, and longitude. We will also touch on the capabilities of parallel processing using another Python library calleddask. -
Discussions
This session will have discussions on how to collaborate on SCINet, desired training topics and collaboration opportunities in the upcoming year, and how someone can contribute to the working group.