AI for Data Analysts
Your SQL or Python skills get you the answer. This module gets you there faster, cheaper, and with proof you can defend to a stakeholder.
Every analyst now works alongside AI. The ones who get ahead aren't the ones who use it the most — they're the ones who know exactly what it's good for, catch it when it's wrong, and can prove their numbers hold up. That's what this module gives you. You'll take a real question, answer it against a genuine 95-million-row dataset, and carry that answer through six lessons: querying faster, working cheaper, auditing for bias, meeting regulatory obligations, presenting to a stakeholder, and publishing a portfolio piece you can actually show an employer. By the end, you won't just have used AI — you'll have a body of work that proves you know how to.
What you'll walk away able to do
| Skill | What it looks like on the job |
|---|---|
| Query faster, without guessing | Get to a working answer in an afternoon on a dataset you've never touched before, instead of a week of trial and error |
| Prove your numbers | Show a real before/after cost comparison and a checked-against-source finding, not "the AI said so" |
| Catch what AI gets wrong | Spot bias, overstated claims, and regulatory exposure before they reach a decision-maker — the judgment that actually protects your job |
| Present like it matters | Turn a finding into a deck and a portfolio page a stakeholder or employer would actually want to see |
Lessons
| # | Lesson | What you'll do | Tool | Time |
|---|---|---|---|---|
| 1 | Where AI Saves You Time | Lock the question you'll answer all module, set up your Claude workspace, and feel the time saved on your very first task | Claude | 4 hrs |
| 2 | Where AI Fills Your Knowledge Gap | Get trustworthy answers from a 113-field glossary you've never read, before you write a single query | NotebookLM | 2.5 hrs |
| 3 | Working Cost-Effectively with AI | Do the same analysis the wasteful way, then the sharp way, and see the savings on your own Usage panel | Claude | ~2.8 hrs |
| 4 | Ethics and Responsible AI Use | Audit your model for hidden bias, then classify its legal risk tier under the EU AI Act with real citations | Claude NotebookLM | 3 hrs |
| 5 | Presenting Insights with AI | Turn your findings into one decision-ready deck a stakeholder can actually act on | Gamma | 3 hrs |
| 6 | Your AI-Powered Brand and Portfolio | Publish a live portfolio page and LinkedIn post built on your finding — and catch AI's own overstated claims before they go live | Claude | 5 hrs |
What will you have by the end?
Six real deliverables — the kind of work a stakeholder or an employer would actually want to see:
| Deliverable | What it contains |
|---|---|
| Your Claude Project | Your locked question and dataset context, ready to use across every AI-assisted step |
| Field Brief | A verified map of which fields answer which of your sub-questions, with the ambiguous ones flagged |
| Results Brief | Your sub-questions answered cost-effectively, with real before/after Usage panel proof |
| Bias Audit & Classification Memo | A proxy-variable bias audit plus a cited EU AI Act risk-tier classification |
| Stakeholder Deck | A presentation-ready narrative built from your Results Brief, with a closing audit-trail slide |
| Published Portfolio Page & LinkedIn Post | A live page featuring your finding, plus a LinkedIn post — checked for overstated claims before you publish |
Frequently asked questions
Is this a tutorial on how to use Claude, NotebookLM, or Gamma?
No. You'll learn when AI actually earns its place in your workflow and when it doesn't — locking a question, closing a knowledge gap, working cost-effectively, auditing for bias, presenting findings, and building a portfolio.
What dataset will I work with?
Real Fannie Mae Single-Family Loan Performance Data, 2003Q1 — roughly 95 million rows across 113 fields. You register on Data Dynamics and download the file yourself.
Do I need a coding background for this module?
Yes — this module assumes your SQL or Python foundation is already in place. Postgres/SQL is the primary worked example throughout; pandas works too. AI accelerates and verifies your querying, it doesn't replace understanding it.
What AI tools will I use?
Claude, NotebookLM, and Gamma — each matched to the task it's actually good at: reasoning and code with Claude, deep-document Q&A with NotebookLM, and narrative decks with Gamma.
What will I walk away with?
A question you answered cost-effectively and in writing, a bias audit and regulatory classification memo, a stakeholder-ready presentation deck, and a live published portfolio page with a LinkedIn post.
How long does this module take?
Around 20 hours across six lessons. The last lesson alone budgets 4 hours, because you're building and publishing a real portfolio page, not a mock exercise.
Contact
Want to Join?
Have questions? We’re here to help! Whether you’re curious to learn more, want guidance on applying, or need insights to make the right decision—reach out today and take the first step toward transforming your career.