alive-analysis

alive-analysis Skill

Data analysis workflow kit based on the ALIVE loop. Provides structured analysis methodology for data analysts and non-analyst roles.

---

Overview

alive-analysis helps structure data analysis work using the ALIVE loop: Ask → Look → Investigate → Voice → Evolve

It serves two personas:

• Data analysts: Deep, systematic analysis with full ALIVE flow
• Non-analyst roles (PM, engineers, marketers): Quick analysis with guided framework

---

ALIVE Loop Reference

Stage 1: ASK (❓)

Core question: What do we want to know — and WHY?

Purpose:

• Define the problem clearly and confirm the requester's REAL goal (not just what they said)
• Frame the question: Is this about causation ("Why did X happen?") or correlation ("Are X and Y related?")?
• Set success criteria and scope boundaries
• Build a hypothesis tree before touching any data
• Set up multi-lens perspective: macro (market/industry) → meso (company/product) → micro (user/session)

Hypothesis Tree

Before diving into data, structure thinking:

Main question: "Why did D30 retention drop?"
├── Internal factors
│   ├── Product changes (releases, feature removals)
│   ├── Channel mix changes (acquisition source shift)
│   ├── Cross-service impact (did another service change affect this?)
│   └── Pricing / promotion changes
├── External factors
│   ├── Seasonality / holidays
│   ├── Competitor actions
│   ├── Market / economic shifts
│   └── Platform changes (iOS/Android policy, algorithm updates)
└── Data artifacts
    ├── Tracking changes (instrumentation broke?)
    ├── Definition changes (metric recalculated?)
    └── Population changes (new user mix shifted?)

Causal vs Correlational Framing

Ask explicitly:

• "Are we trying to prove X caused Y? Or just that they move together?"
• "If we find a correlation, what would we need to prove causation?"
• This determines the methodology: correlation → observational analysis; causation → quasi-experimental or controlled experiment