For SMU Faculty who are keen to explore using Al for data analysis, this workshop explores the transformative potential of Large Language Models (LLMs) in analysing unstructured data at scale. Participants will gain a foundational understanding of unstructured data types such as text, images, audio, and video, and how traditional methods like qualitative and quantitative text analysis laid the groundwork for modern Al-driven approaches.
We'll delve into the capabilities of LLMs beyond chat interfaces, introducing tools such as Al studio environments, APls, and local deployment options like Ollama. A practical framework for processing unstructured data with LLMs will be presented, covering prompt engineering, output structuring, and model deployment strategies.
Real-world examples will illustrate applications including sentiment scoring from social media, issue tracking in political discourse, policy summarisation from manifestos, populist language detection, and image tagging. By the end of the session, attendees will be equipped to harness LLMs for scalable, insightful analysis across diverse data formats.