Get started with SEA by visiting the SEA main page.
Click to expand the examples.
Start a new SEA conversation using the "launch" button, or copy/paste prompt in SEA (opens in a new tab), or view examples of past SEA conversations that began with these prompts.
Explain data to me. I am new to researching sea level and tidal information. Share information about the data available at UHSLC and within this app. Discuss ways to interact with the data for the selected station, such as requesting to see samples, perform analyses, or download files. Ask me what I would like to explore next.
Launch conversation View exampleAnalyze data for me. I am new to researching sea level and tidal information. Briefly explain what data is available for the selected station. Jump into an example, such as using the Fast Delivery daily data to calculate monthly means and then calculating the trend (remember there could be data gaps that should be ignored when calculating trends). Ensure that trend calculations are clearly specified as per year. When calculating trends, always verify the time unit and convert to an annual rate, if necessary, before presenting results. Show your code and plot the results. Offer to calculate monthly anomalies by subtracting the annual cycle climatology. Ask me what I would like to explore next.
Launch conversation View exampleConvey information to me. I am especially interested in understanding tidal datum information. Teach me about the tidal datums, especially about how they relate to coastal impacts and how to convert data between different reference datums. Jump into an example for the selected station by loading the tidal datums and plotting them as both relative to Station Zero and MHHW datums. Show your equations, code, and plots. Ask me what I would like to explore next.
Launch conversation View examplePropose research ideas to me. I am looking for creative new ideas to research using the sea level and tidal data at UHSLC. Briefly explain what data is available. Outline several neat research questions to explore. Pick one question and help me get started with the analysis for the selected station. Show your code and plot the preliminary results. Ask me if there is a research topic that I would like to further explore.
Launch conversation View exampleAsk questions that interest you. SEA will help focus the discussion on the scope of what it can assist with.
The application is designed to help users analyze and understand data from UHSLC.
Sure, upload your data as long as it relates to tides and sea levels (up to 10 files per conversation that are smaller than 10MB/each; .csv, .txt, .json, .nc, .xlsx, .tif).
Safety and security features include: No tracking of user web browsing. Code executions occur on UHSLC computers in a partitioned environment separate from the database.
Mistakes happen! Always check responses, and let SEA know if something doesn’t seem right. You can continue the work on your own or start fresh with a new conversation.
Yes! The project developers, Matthew Widlansky and Nemanja Komar, are happy to hear your feedback and provide suggestions.
SEA is powered by the GPT-4o large-language model from OpenAI, with access to Station Explorer data from UHSLC, usage instructions, and a computing environment for analysis. SEA performs best when users ask for help analyzing and understanding sea level and tidal data.
The preprint publication detailing the Intelligent Data Exploring Assistant (IDEA), which SEA is based on, is available at:
Building an Intelligent Data Exploring Assistant for Geoscientists
Cite as: Matthew J. Widlansky, Nemanja Komar. Building an Intelligent Data Exploring Assistant for Geoscientists. ESS Open Archive. February 24, 2025.
DOI: 10.22541/essoar.174042987.76981404/v1 (Registration in progress)
This is a preprint and has not been peer-reviewed.
The source code and implementation details for an Intelligent Data Exploring Assistant (IDEA) like SEA are available on GitHub:
utils
directory)
Press Roundtable: Chatbot Lab Assistants: Generative AI is Changing How We Do Science, AGU 2024
Station Explorer Assistant for sea level science communication, December 2024
Introduction to Station Explorer Assistant, CIMAR 2024
Introduction to using Large-Language Models for sea level science communication, AGU Ocean Sciences Meeting 2024
For more information, please contact Matthew Widlansky (Principal Investigator) or Nemanja Komar (Lead Developer).