Data Analysis with Generative AI

Course details

Generative AI can analyze a dataset in seconds, but only if you know how to prompt it well and catch it when it's wrong. In this hands-on course, explore how to put tools like ChatGPT to work on real data, including a customer-churn dataset and USPTO trademark records. Learn how to prompt an LLM to explore data, generate summary statistics, and surface insights you can act on. Discover how to clean and standardize messy data, deduplicate and group entity names, and weigh modeling approaches such as logistic and survival models. Find out how to verify that an LLM analyzed your actual file and how to detect hallucinations before they reach your conclusions. By the end of this course, you'll be equipped with the skills you need to use generative AI to analyze data critically and reliably. This intermediate-level course is an ideal fit for data analysts, data scientists, and anyone eager to analyze data with AI.

Learning objectives
Prompt an LLM to explore and summarize a dataset.
Generate useful summary statistics for a business context.
Clean and standardize messy data such as entity names.
Detect and avoid AI hallucinations in analysis.
Use AI to help guide model selection.
Verify that AI output reflects your actual data.

Concepts

Introduction

Course overview and setup
GenAI and its role in data analysis
Crafting effective prompts for data analysis

Data Workflow and Collection

Structured vs. unstructured data
Collecting and organizing data for analysis - Part I
Collecting and organizing data for analysis - Part II

Next-Level Data Cleaning with GenAI

Data handling
Automating data cleaning
Dataset reduction and transformation
Feature selection and engineering

Data Analysis to Extract Meaningful Insights

Building a model using GenAI
Improving our model using GenAI - Part I
Improving our model using GenAI - Part II

Hallucinations, Ethics, and Wrap-Up

Ethical implications of GenAI and responsible use