Livebook + Elixir - Expanded Project Guides
Generated from: LEARN_LIVEBOOK_ELIXIR_DEEP_DIVE.md
This folder contains twelve project-level deep dives for learning Elixir through reproducible, interactive, observable, and deployable Livebook workflows.
Overview
The series begins with one deterministic .livemd report, then adds Kino forms, boundary validation, supervised processes, Explorer pipelines, safe SQL, live dashboards, cloud file references, a custom Smart Cell, numerical inference, remote runtime diagnostics, and a deployed multi-session app. Every project requires observable evidence and a clean-runtime replay. Projects that touch operational systems remain read-only, use disposable training targets, and state their trust boundary.
Project Index
| # | Project | Difficulty | Time | Key Focus |
|---|---|---|---|---|
| 1 | Executable Energy Field Report | Level 1 | 5–8 hrs | .livemd, Elixir values, clean replay |
| 2 | Interactive Capacity Planning Workbench | Level 1 | 6–10 hrs | Kino forms, validation, client semantics |
| 3 | Messy CSV Triage Clinic | Level 2 | 10–16 hrs | parsing, tagged errors, redaction |
| 4 | Fault-Tolerant Telemetry Lab | Level 3 | 16–24 hrs | processes, supervision, backpressure |
| 5 | City Mobility Data Pipeline | Level 2 | 12–20 hrs | Explorer, lazy plans, Parquet |
| 6 | Safe SQL Incident Investigation Runbook | Level 2 | 12–20 hrs | read-only SQL, secrets, audit evidence |
| 7 | Interactive SLA Command Center | Level 3 | 18–28 hrs | VegaLite, responsive events, freshness |
| 8 | Portable Cloud Data Lake Lab | Level 3 | 16–26 hrs | file references, S3, portability |
| 9 | Tested Domain Smart Cell Toolkit | Level 3 | 20–30 hrs | Kino.JS.Live, source generation, tests |
| 10 | Nx + Bumblebee Image Triage Lab | Level 3 | 20–32 hrs | tensors, serving, human override |
| 11 | Safe Attached-Runtime Triage Console | Level 3 | 20–30 hrs | runtime boundaries, distributed Elixir |
| 12 | Deployed Multi-Session Operations Decision App | Level 4 | 35–50 hrs | app deployment, identity, operations |
Dependency Graph
P01 Reproducibility + Elixir
├──▶ P02 Kino forms
├──▶ P03 data contracts
│ ├──▶ P05 Explorer pipeline ──▶ P08 cloud data
│ └──▶ P06 safe SQL ───────────┐
└──▶ P04 OTP telemetry ──▶ P07 SLA dashboard
│ │
P02 ─────────────────────┘ ├──▶ P09 Smart Cell
P05 ─────────────────────────────────┤
P05 + P07 ──▶ P10 Nx image triage │
P04 + security primer ──▶ P11 runtime console
│
P06 + P07 + P08 + P09 + P10 + P11 ──▶ P12 deployed app
The graph shows recommended dependencies, not hard blockers. Experienced Elixir engineers may start at Project 4 or 9; data practitioners may move from Project 3 directly to Project 5.
Learning Paths
Elixir Beginner
P01 -> P02 -> P03 -> P04 -> P05 -> P07 -> P12
Data and Analytics
P01 -> P03 -> P05 -> P06 -> P07 -> P08 -> P10 -> P12
Elixir Application Engineer
P01 -> P04 -> P07 -> P09 -> P11 -> P12
Internal Tools / Solution Architect
P02 -> P06 -> P07 -> P08 -> P11 -> P12
Numerical and ML
P01 -> P03 -> P05 -> P07 -> P10 -> P12
Prerequisites
- Livebook Desktop or a compatible self-hosted Livebook
- Basic programming and terminal literacy
- Git for reviewing .livemd diffs
- Two browser windows or profiles for collaboration tests
- Docker for Project 12
- A disposable PostgreSQL or SQLite database for Project 6
- A disposable named Elixir/Phoenix node for Project 11
- Optional local S3-compatible storage for Project 8
Livebook Desktop is the simplest starting environment. If installing Livebook directly through Elixir, follow the current official requirements rather than copying old version pins from a project transcript.
How to Use These Guides
- Read the parent guide’s matching theory chapters first.
- Complete the Core Question and Thinking Exercise before implementation.
- Work through the three implementation phases and save checkpoint evidence.
- Use layered hints only after writing a concrete hypothesis.
- Run critical tests, failure cases, and two-client checks where required.
- Reconnect the runtime and perform a top-to-bottom replay.
- Record lessons learned and one design change before marking completion.
Shared Safety Rules
- Never commit credentials, cookies, tokens, private URLs, or personal data.
- Use seeded training data and disposable targets before connecting to any real system.
- Keep Projects 6, 11, and 12 read-only.
- Treat Livebook authoring access as code-execution access.
- Treat notebook stamps as remembered resource authorization, not a sandbox.
- Bound rows, bytes, time, concurrency, event history, and model inputs.
- Preserve provenance and label live, snapshot, cached, stale, partial, and failed results accurately.