In research, the ability to “speak” directly to the data not only simplifies analysis but also democratizes the use of AI—allowing anyone, regardless of coding expertise, to uncover meaningful insights. This talk will showcase a cutting-edge Shiny for Python application that seamlessly integrates interactive data exploration with AI assistance. Developed using LangChain, pandas, scikit-learn, and statsmodels, the application empowers researchers to upload datasets via DOI links from the International Potato Center (CIP) Dataverse. Once uploaded, users can select specific tables, engage with a conversational DataFrame agent, and instantly generate visualizations and statistical analyses—without writing code.

Beyond interactive exploration, the application features a streamlined workflow for reporting. Users can compile custom Quarto documents that include AI-generated plots, user-created word clouds, conversation transcripts, and table outputs. By blending an intuitive interface with AI-driven data analysis, we’re illustrating how Shiny for Python can serve as a powerful tool for life sciences and beyond. Attendees will learn practical strategies for building similar AI-assisted Shiny apps—from handling complex datasets to ensuring reproducible reports—offering a glimpse into the future of accessible, data-centric research.