A self-study curriculum, structured like the ratio that built it
Zero to IBM‑Certified
Data Scientist
Ten phases. One ratio. Every phase of this path is sized against φ = 1.618 — the same proportion that governs a nautilus shell and a Renaissance canvas — so the curriculum compounds the way the spiral does: slow at the centre, wide at the edge.
50+ hand-checked free resources · 10 phases · 1 destination certificate
The curriculum
The Path, in ten turns of the spiral
Each phase narrows in scope and deepens in skill — the way each new golden rectangle is smaller than the last but holds the same proportion. Work top to bottom; nothing here skips ahead of what it needs.
Everything in one shelf
The Resource Library
Every link from every phase, in one searchable shelf. Filter by what kind of learning you're in the mood for — reading, watching, building, or referencing.
No resources match that search. Try a shorter word.
Where the path leads
Mapping to the IBM Data Science Professional Certificate
This path is sequenced to mirror IBM's own programme on Coursera — now expanded to a twelve‑course track that adds Generative AI and a dedicated career‑and‑interview course on top of the original Python, SQL, analysis, visualization, machine learning, and capstone line‑up. Free-study each phase here first; the certificate then formalizes and badges what you already know.
View the certificate on Coursera ↗| Codex phase | IBM course it corresponds to |
|---|
IBM's course list and pricing change over time — the table above reflects the structure as of early 2026. Always check the live Coursera page before enrolling.
How to actually use this
Begin at 1, stop at nothing
One phase at a time
Don't open phase 4 because phase 6 looks more exciting. The order is load-bearing — SQL leans on Python, machine learning leans on the math.
Build, don't just read
For every course or article, ship one small artifact — a notebook, a query, a chart. Push it to GitHub. The portfolio is the actual certificate.
Re-read the Wikipedia page last
Skim it first for orientation, then return after the hands-on work — it will suddenly make complete sense instead of half sense.
Let the capstone be real
Pick a dataset you actually care about for your final project. Care is the only thing that survives the boring debugging hours.