Integrating Large Language Models into Economic Research and Data Analysis with tidyllm

Large language models (LLMs) are transforming how we process and analyze text, images, and structured information. I developed a simple and user-friendly R package to use both local LLMs and LLM APIs efficiently in different typical data workflows, enabling seamless integration of these models into research tasks.

This talk will present and demo the package and explore three key use-cases in economic and social science research:

Text and Image Classification: Automate the classification of open survey responses, images or whole documents.

Structured Information Extraction: Convert unstructured text from research papers, legal contracts, or financial reports into clean structured datasets for further analysis.

Semantic Search and Document Similarity: Use embedding models to find relevant documents, or match similar texts across large text corpora.