Get Updates

AI Tools for Data Analysis: From Chatbot to AI Agents - A Short Course

An 8-Hour Livestream Seminar Taught by Boris Nikolaev, Ph.D.

Read reviews of this course
Download Sample Course Slides

This seminar is part of the AI-Enabled Data Analytics Certification, a series of 4 expert-led courses designed to build practical AI skills for research and data analysis. Contact us to learn how to complete your certification and access special pricing.

 

In 2026, AI beat every human at the world finals of competitive programming and helped disprove a math conjecture that had stood for 87 years. The same tools can now handle many of the technical tasks involved in data analysis. They pull the data, clean it, run the models, draw the figures, and build interactive data stories. Most of us are only scratching the surface.

In this course, you will learn to delegate work to AI at three levels: individual tasks, entire projects, and multistep processes. We work with Claude and ChatGPT, including their coding-agent tools, Claude Code and Codex. You will also set up an AI agent that runs your statistical software of choice, such as Stata, R, or Python, and keeps a trail of its work.

The first day focuses on using AI tools for individual analytical tasks, while the second explores how AI coworkers and agents can manage larger projects and automate multistep workflows. Throughout, we will encourage creativity, practice designing checks that catch silent errors, and develop the judgment needed to evaluate what a result can and cannot claim.

By the end of the course, you’ll be able to use AI tools to support data analyses and create websites, dashboards, interactive presentations, and other outputs without writing code yourself. Some sample projects we will build include a 3D dashboard like the Global Happiness Atlas and a data story like The Name Market.

Now, the only limit is our imagination.

For additional examples, a detailed curriculum, and setup instructions, visit the instructor’s course website.

Starting February 25, this seminar will be presented as an 8-hour livestream workshop via Zoom. Each day will feature two lecture sessions with hands-on exercises, separated by a 30-minute break. Live attendance is recommended for the best experience. If you can’t join in real time, recordings will be available within 24 hours and accessible for four weeks after the seminar.

Closed captioning is available for all live and recorded sessions. Captions can be translated to a variety of languages, including Spanish, Korean, and Italian. For more information, click here.

ECTS Equivalent Points: 1

More Details About the Course Content

Live and hands-on. We demonstrate using Claude, and you follow along using either Claude or ChatGPT.

Real datasets from the start. We use panel data on world happiness (165 countries since 2005), combined with live data from the World Bank and other sources such as FRED. We focus on real examples that can be messy.

Your own website, for free. You will learn to create and publish websites at no cost, so you can promote your own work. Everything we build goes online with its own link, and you leave with work you can show, not just notes.

What you take home. Every prompt from class, the skills we write together, and all the projects and workflows, ready to reuse on your own data.

Computing

You need one paid AI plan, Claude Pro or ChatGPT Plus (about $20 a month; you can cancel after the course), and its desktop app. Live demonstrations will use Claude. You can use ChatGPT and Codex for comparable workflows, although some features and steps may differ. The desktop app matters on day two, when the AI works directly from folders on your own computer.

You will also need a free FRED API key to retrieve economic data on day one and a free Vercel account to publish dashboards and websites on day two. You do not need Python, R, additional statistical software, a GitHub account, or a powerful computer.

Setup takes about 30 minutes. Step-by-step instructions, with a checklist that saves your progress and a two-minute test drive, are on the course website’s setup page. Paid plans have usage limits, so please go easy on heavy use right before each session.

Given the rapidly changing AI landscape, platforms and API access requirements may change before the course begins. We will provide up-to-date setup guidance before the seminar, including recommended AI tools, account requirements, and any estimated API costs for course exercises.

Who Should Register?

This course is for anyone who works with data, already uses ChatGPT or Claude, and suspects there is a lot more. It suits researchers, graduate students, faculty, and analysts in industry, government, health care, and nonprofits. It works whether you are new to data analysis or have run models for years.

No coding experience is needed. If you already work in Python, R, or Stata, you will learn to hand off the routine work and spend your time on the questions and the checks. If you have never written code, you will get real results in plain English.

Interested in broader research applications? Explore our Using LLMs for Social Science Research seminar.

Outline

Day one: delegating tasks (the chatbot)

    • Where we actually are: how capable the tools have become, what benchmarks measure, a first analysis in a minute
    • Models, modes, and apps: fast and thinking models, deep research, code execution, artifacts you can share, Claude and ChatGPT side by side
    • Prompting: a four-part checklist (context, goal, output, done), asking for the code, iterating on an answer
    • Setting up a project: a folder, its data, and standing instructions that carry across sessions
    • Getting data: via FRED and World Bank APIs, CSV, Excel, and PDF tables, merging sources
    • Cleaning and reshaping: missing values, outliers, duplicates, recoding, new variables
    • Describing the data: summary and balance tables, group breakdowns, correlation matrices, exporting to Word, tables in LaTeX
    • Figures and maps: trends, rankings, distributions, scatter plots with fit lines, world maps
    • Regression: OLS, controls, fixed and random effects, regression tables
    • Beyond the regression table: mediation, structural equation models, using custom packages
    • Going further: latent profile analysis and event studies (if we have time)
    • A first dashboard: an interactive dashboard others can open and share

Day two: delegating projects and processes (the coworker and the manager)

    • Folder, goal, permissions: connecting Claude Code or Codex to a project folder, setting permissions, and reviewing its plan
    • The toolkit: CLAUDE.md house rules, project memory, skills, sub-agents, MCP connectors, plugins
    • What we build: a full report, a website, a multiverse analysis, and data stories
    • Designing the loop: moving from a single AI coworker to a team of agents, with defined goals, verification checks, and stopping rules
    • Verification design: silent errors, writing the check down, reviewer agents
    • Judgment and taste: evaluating what a chart can and cannot claim and documenting standards as a reusable skill
    • The closing hour: Is AI making us stupid? What is left for experts? What would we never delegate?

Reviews of AI Tools for Data Analysis: From Chatbot to AI Agents

“The course was eye-opening, and Boris was very knowledgeable and at the cutting edge of the AI juggernaut. His slides and videos were very informative.”
  Ralph Hosoki, Sophia University

“It was clear that Boris has a lot of experience working in Claude and ChatGPT. I also liked the extra tutorials that we can do outside of class with a deeper dive into specific components of Claude. His passion for working in AI was clearly evident during the class!”
  Edie Weller, Boston Children’s Hospital

“The diversity in applications for using ChatGPT for data analysis, especially the creation of interactive dashboards and HTML to host on a website, was invaluable.”
   Christine Davis, The National Center for Construction Education and Research

“I have been using ChatGPT every day since 2023 for research, teaching, and even creative writing, though primarily to code advanced statistical models. Despite this experience, I learned so much new information about what AI can do and its potential.”
  Jane C. Daquin, Sam Houston State University 

“I think Boris explained all the topics very well. He knows a lot and was really interactive. Data analysis and AI are very, very important topics that really help our academic careers.”
  Laura Fabregat, Universitat Rovira I Virgili

Seminar Information

Thursday, February 25 –
Friday, February 26, 2027

Daily Schedule: All sessions are held live via Zoom. All times are ET (New York time).

10:30am-12:30pm (convert to your local time)
1:00pm-3:00pm

Payment Information

The fee of $695 USD includes all course materials.

PayPal and all major credit cards are accepted.

Need funding from your employer or institution? Use our funding request template.

Our Tax ID number is 26-4576270.