Applied AI for Data Analysis and Automated Workflows

(0 Credits)

The objective of this workshop is to provide an overview of using LLMs as practical tools for data collection and analysis, and allowing students to create a repeatable workflow to collect, clean and analyze data. The workshop will initially focus on the basics of how LLMs operate, and how effective prompt engineering can reduce the margin of error and increase productivity when executing specific tasks. With a focus on QAing and translating subject area expertise into detailed instructions, the goal is for students to gain experience with using AI for general tasks, as well as for initial data cleaning. After introducing prompt engineering, the workshop will focus on how to leverage AI in order to have access to technical data sources without prior knowledge of programming. The workshop will cover the tools that now let AI connect to external data, such as APIs and MCP (Model Context Protocol) connectors, and give participants the chance to use them directly. Ultimately, students should gain experience asking AI to work with these sources and extract large amounts of data that can later be used for research. The workshop will then shift to creating repeatable workflows, and configuring specific instances of the models to specialize in one task. Students will learn how to turn AI from a chatbot into a highly specialized assistant that can automate tasks in the realm of data collection and analysis (such as using the web to collect data and perform light analysis). Finally, the workshop will close with an overview of how workflows can be made available beyond a single person's account. The options discussed will include sharing a workflow's instructions with a colleague, deploying it across multiple people within the same organization through a shared Claude or ChatGPT workspace, and publishing it as a standalone application that anyone can use. As a non-credit workshop, there will be no final grade assigned. However, by the end of the workshop, students will be expected to develop a set of project instructions that turns Claude or ChatGPT into a guided, highly specialized assistant capable of using the web to collect data and perform light analysis.