Skip to content

ai-analyst-lab/ai-analyst-starter

Repository files navigation

AI Analyst (Starter)

Welcome to Day 1 of the AI Analyst Bootcamp. This is your starting repo -- nearly empty on purpose. You'll build your own AI analyst from the ground up.

Prerequisites

Setup

1. Clone this repo

git clone https://github.com/ai-analyst-lab/ai-analyst-starter.git
cd ai-analyst-starter

2. Install Python dependencies

pip install -r requirements.txt

3. Generate the dataset

pip install -r data-generation/requirements.txt
python data-generation/generate.py

This creates ~8M rows of realistic e-commerce data in data/practice/ -- users, orders, sessions, clickstream events, experiments, support tickets, and more.

4. Open Claude Code

claude

Claude reads CLAUDE.md and becomes your AI Product Analyst. Try asking it a question:

"What's the conversion rate by device type?"

"Which acquisition channel brings the most valuable users?"

"How does NPS differ between Plus members and non-members?"

What's Here

CLAUDE.md                    ← The persona (read this first)
data-generation/             ← Scripts to generate practice data
.claude/skills/              ← 3 starter skills (kickoff, show-off, question-framing)
agents/                      ← Empty -- you'll build your first agent here
.knowledge/                  ← Dataset schema and connection config
themes/                      ← Marp slide themes
templates/                   ← Slide deck template

What's NOT Here (Yet)

No pre-built agents. No pipeline. No helpers. No chart library. You build what you need.

That's the whole point of Day 1: understand how the pieces work by building them yourself.

Key Concepts

CLAUDE.md -- The persona file. This is what turns Claude Code from a general assistant into a domain-specific analyst. Everything Claude knows about its role, data, and rules comes from here.

Skills (/.claude/skills/) -- Instruction files triggered by slash commands or context. Read by Claude at runtime. Think of them as "how-to guides" for specific tasks.

Agents (/agents/) -- Markdown templates with {{VARIABLE}} placeholders. Define reusable analytical workflows. Read the template → substitute variables → execute.

Next Steps

  1. Read CLAUDE.md to understand the persona
  2. Try /kickoff to introduce yourself to the community
  3. Ask Claude an analytical question about NovaMart
  4. Build your first agent (check agents/CONTRACT_TEMPLATE.md for the format)
  5. Try /show-off to share what you built

About

AI Analyst Starter — Day 1 bootcamp repo

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors