TutorialsFundamentals & Core Concepts
See the finished product, then understand the 5-stage workflow, 4 core concepts, and 6 common scenarios. Build the core mindset: AI handles implementation, you handle the thinking.
The Finished Product: What You'll Build
The course opens by showing the final deliverable: an A4 one-page management dashboard + HTML analysis report. March's sales decline is immediately visible, top products and regions are clearly ranked, with 3 actionable recommendations.
This isn't a demo — it's what you'll build with your own hands in the 9 steps that follow. When step 9 is done, the report is finished, bringing you right back to this opening.
What Can Be Done This Way
Sales Analysis
Total revenue, trends, product/region rankings — auto-read detail, calculate metrics, generate charts and report.
Expense Analysis
Department cost rankings, expense type proportions, anomaly detection (flagged for review, not flagged as violation).
User Behavior Funnel
Conversion rates at each stage, biggest drop-off points, hourly activity patterns.
Satisfaction Surveys
Per-question averages, lowest-scoring dimensions, sentiment analysis + keyword extraction.
Report Consolidation
Multi-table reconciliation, monthly rollups, batch reading with auto-verification.
Chart Styling/Dashboards
Specify a style, generate standardized charts, and assemble a one-page management dashboard.
Data Analysis Workflow: 5 Stages
The complete data analysis pipeline has 5 stages. The 9-step workflow that follows is an expansion of this pipeline:
| Stage | What It Does | Who's Responsible |
|---|---|---|
| Data | Prepare, clean, verify fields | You |
| Analysis | Calculate metrics, slice dimensions | AI |
| Charts | Select chart types, generate visualizations | AI |
| Report | Distill findings, write recommendations | AI draft + You approve |
| Action | Judge whether recommendations are executable | You |
4 Core Concepts
Metrics & Definitions
Metrics = what numbers to use (total revenue, average order value, achievement rate). Definition = how that number is calculated (e.g., "order count" — by unique order ID or by row count?). Whether AI calculates correctly depends on whether the definition is clear.
Analysis Dimensions
Which angles to slice the data by: time, product, region, customer type. The same dataset, different questions, completely different dimensions.
Human-AI Division of Labor
AI handles implementation (running data, calculating metrics, generating charts, building reports). You handle the thinking (defining metrics, choosing dimensions, checking results, making decisions).
Data Analysis Workflow
Data → Analysis → Charts → Report → Action. The 9-step workflow in this course is an expansion of this pipeline.
Core Principle
AI lowers the barrier to data analysis, but the "analytical thinking" is still your job. AI handles implementation; you handle the thinking — define metric definitions, explain field meanings, choose the right questions, verify results, and judge whether recommendations are actionable.
Think of AI as a very smart subordinate. As the boss, your job is to assign the work clearly and make the final call. In the next lesson, we'll learn how to assign that work — the prompt formula.
Data Analysis with AI
9-step workflow, 3 hands-on exercises, 8 prompt templates — turn a spreadsheet into an analysis report in 3 minutes without SQL or Python.
Prompt Formula & 6 Background Elements
One universal formula (Role + Task + Background + Format + Constraint) plus 6 data-specific background elements to stop AI from guessing and reduce rework.