Build science apps with AI

SCIWS14 · Scientific Workshop

Using AI to Build Science-Based Apps with Vibe Coding

Transform the way that science is done or redefine how people engage with earth and space data by learning how to vibe code your own software apps. Vibe coding is a powerful approach for programming that utilizes AI agents to create your own fully-functioning software. This workshop will introduce you to tools, strategies, and resources that can be used to create your own software even if you have no programming experience. Regardless of your experience level, you will leave this workshop having made a fully functioning app and ready to explore new ways of collecting, analyzing, sharing, or storytelling about the data you use in your science!

The workshop will focus on the interface between defining real-world problems and prototyping fast software solutions to explore and communicate those problems. We will not only introduce tools you can use to build your own app, but also the important role that vibe coding can play in building shared mental models around complex problems. On the technical side we will introduce you to how large language models (LLMs) work and can be used to generate code, how to access and use science data to power your software app, and ways that you can use AI to make your science accessible to a broader audience. We welcome all experience levels to participate in the workshop because innovation happens where ideas, experience, and passion meet! Whether you want to get creative about how we can do science differently, find ways to increase participation in science, or just do something that you think is really cool, this workshop is for you!

TH33I · Town Hall

From Data to Action: Cocreating Sustainable Solutions Through AI, Community-Based Science, and Multiplatform Earth Observation

Community-based participatory science has tremendous potential to engage people in understanding our planet, but Earth observation (EO) data remains largely inaccessible to non-specialists. This town hall explores how artificial intelligence and machine learning can democratize data access across multiple platforms (e.g. satellite imagery, ground sensor networks, citizen-collected observations, and model outputs) enabling community scientists to extract meaningful insights and develop sustainable solutions to local environmental issues. We'll discuss practical pathways where AI tools lower barriers to entry: automated data preprocessing, machine learning models that identify features of interest, and user-friendly interfaces and agentic workflows that translate complex information into actionable knowledge. Critically, we emphasize cocreation - designing AI tools collaboratively with communities rather than imposing top-down solutions. Examples span diverse applications: land use monitoring, water quality, vegetation health, and air quality assessment.

We'll address key challenges: ensuring data quality and scientific rigor, training requirements and model biases, perceptions and fears around the use of AI, integrating community findings into research and policy, and achieving equitable access and benefit-sharing. Corporate technology partners play a key role alongside researchers and agencies in building scalable, sustainable platforms that serve diverse communities. The Town Hall will draw on discussion from a panel of experts to inspire real-time engagement with the audience and collaboratively create a shared vision for the future. Panelists will represent diverse perspectives: AI and machine learning researchers, community science coordinators, data providers, corporate technology partners, and active community scientists. We'll discuss funding, institutional partnerships, and corporate engagement models that support human-centered AI for EO.

This town hall will produce a report catalyzing a research agenda that connects community science, multi-platform data, and machine learning to create equitable, sustainable pathways for public engagement in environmental science globally. Note: This Town Hall will be part of a series of inter-related activities at AGU.

SE51A · Science Exchange

Software Showcase: Innovating Data Collection, Analysis, and Communication in the Earth & Space Sciences

This Science Exchange session will showcase innovative software strategies for collecting, analyzing, or sharing data that supports discovery and community engagement in the Earth and Space Sciences. Participants in the showcase will share products, prototypes, and concepts that could transform the way that scientific data are used within and outside research settings - including contributions relevant to operational, science communication, and participatory science applications. Let your creative juices flow and share your vision for how science can be performed or shared to increase the impact of the Earth and Space Sciences on society. This showcase will be the culmination of the Hack the Earth Software Hackathon event at AGU26, but all contributions are welcome regardless of participation in that event.

All products shared in this Science Exchange Showcase must be submitted to this site as a public URL no later than the day before the start of the session. Submissions can be for a live web application, webpage, or other document that allows you to share the working product or conceptual idea that you would like to showcase. All submissions will be open for public comment and the highest ranking contributions will be invited to share their work live during this session at AGU26.