Loading…
2025 January ESIP Meeting
We are a home for Earth science data and computing professionals. Our sessions bring together the community for hands-on, interdisciplinary deep dives as we explore "Innovation to Impact" this year. Learn more about ESIP: esipfed.org

to bookmark your favorites and sync them to your phone or calendar.

Venue: Room 4 clear filter
Tuesday, January 21
 

1:30pm EST

Enhancing reproducible scientific workflows with the R ‘targets’ package
Tuesday January 21, 2025 1:30pm - 3:00pm EST
Writing code has become an integral component of conducting scientific research, especially as datasets size and complexity have grown. Scientists use code to download and clean data, prepare visualizations, calculate statistics, run models, and more. Just as the use of code in environmental research has grown, so has the ecosystem of tools and techniques for building robust analyses. Libraries in the programming language R have expanded to meet the needs of a growing user base in the scientific community, particularly through the open science community, rOpenSci. In this session, we will focus on a particular package in the rOpenSci ecosystem called ‘targets’, which enables users to build robust, data pipelines that enable reproducible and efficient scientific workflows. We will introduce the concepts of dependency tracking that underpin the package, host an interactive demo to build a small pipeline using ‘targets’, and share a few examples of ‘targets’ pipelines built for large research projects. Attendees should leave this session feeling inspired and equipped to begin constructing data pipelines using ‘targets’ for their own projects.

Value to Session Participants: Session participants will leave with an example of a reproducible workflow, and practice writing and running code with dependency management enabled. This should give them a starting point for future projects that can leverage these techniques.

Recommended Ways to Prepare: Skim the homepage of the ‘targets’ rOpenSci docs to understand the high-level summary and philosophy of this approach at https://docs.ropensci.org/targets/. Consider watching the 4-minute demonstration video that shows an example workflow. If you are not familiar with R functions, please read about them in this R for Data Science chapter at https://r4ds.hadley.nz/functions.html.
Speakers
Tuesday January 21, 2025 1:30pm - 3:00pm EST
Room 4
 
Wednesday, January 22
 

1:30pm EST

Cloudy with a Chance of Science: Exploring Innovations in Cloud-Native Geospatial Technologies
Wednesday January 22, 2025 1:30pm - 3:00pm EST
Wednesday January 22, 2025 1:30pm - 3:00pm EST
Room 4
 
Thursday, January 23
 

4:00pm EST

AI Agents and KG-Driven NSAI: Driving Innovation in Geoscience Data Analysis
Thursday January 23, 2025 4:00pm - 5:30pm EST
This session explores AI applications in geoscience, focusing on automated data analysis workflows and deposit type classification techniques using NSAI guided by KGs. We will demonstrate how AI agents leverage open LLMs to perform mineral data analysis and showcase a neuro-symbolic approach that embeds knowledge graphs of deposit rules within neural networks to enhance mineral deposit classification accuracy.

Value to Session Participants: Participants will gain insights into scalable AI-driven workflows and the integration of NSAI guided by KGs to enhance geochemical-based deposit type classification, exploring practical applications for automating data analysis and enhancing classification in geoscience and beyond.

Recommended Ways to Prepare: We will provide a Jupyter Notebook demo for our audience to interact with.
Speakers
JZ

Jiyin Zhang

Research Assistant, University of Idaho
Thursday January 23, 2025 4:00pm - 5:30pm EST
Room 4
 

Share Modal

Share this link via

Or copy link

Filter sessions
Apply filters to sessions.
America/New York