Free Learning Roadmap

How to Become a Data Scientist

Statistics, experimentation, causal inference, and ML applied to product decisions. Communicates findings that change what the business does.

Topics
11
Resources
38
Cost
Free
Most resources are free or freemium
Format
Self-paced

1. Programming

core
Python for Data Science

pandas, NumPy, scikit-learn, matplotlib/seaborn. Notebook fluency plus enough software engineering to ship.

SQL for Analytics

Window functions, CTEs, complex joins. Most of your day is SQL against a warehouse.

2. Statistics

core
Inferential Statistics

Distributions, hypothesis testing, confidence intervals, power. The stuff that stops you from claiming causal effects from noise.

Bayesian Thinking

Priors, posteriors, PyMC. Bayesian methods often give more interpretable answers than p-values for business decisions.

3. Experimentation (A/B Testing)

core
A/B Test Fundamentals

Sample size, MDE, power, sequential testing, novelty and primacy effects. Do experiments right, or don't do them.

When You Can't A/B

Diff-in-diff, synthetic controls, regression discontinuity. For product launches you can't randomize.

4. Machine Learning

core
Supervised & Unsupervised Learning

Regression, classification, clustering, dimensionality reduction. Know when to reach for which.

Gradient Boosting (XGBoost, LightGBM, CatBoost)

Wins most tabular ML problems in production. Understand feature importance, early stopping, hyperparameter tuning.

5. Causal Inference

core
The Causal Framework

Correlation isn't causation. DAGs, counterfactuals, confounders. The most under-taught skill in most stats courses.

6. Data Visualisation

core
Static Visualisation

matplotlib, seaborn, altair, ggplot. Not just to make charts — to reveal patterns you missed.

Dashboards & BI

Tableau, Looker, Metabase, Superset. Analysts get their work productionized through dashboards.

7. ML in Production (Lite)

recommended
From Notebook to Production

Refactor notebook code, add tests, package as a job or service. The bridge from analysis to shipping.

Feature Stores

Feast, Tecton. If your ML models need real-time features, you need one of these.

8. Business Communication

core
Storytelling with Data

Executives don't read notebooks. They read a headline, a chart, and one recommendation.

9. Deep Learning (Optional Depth)

recommended
PyTorch

The default framework for research + production. Learn tensors, autograd, DataLoader, training loops.

Transformers & LLMs

Attention, tokenization, fine-tuning. Even non-NLP DS need conversant literacy here now.

10. Engineering-Adjacent Skills

recommended
Git & CLI Basics

You'll collaborate with engineers on shared codebases. Enough git and bash to not embarrass yourself.

Docker for Reproducibility

Package your notebook environment so someone else can run it in a year.

11. Career

optional
Communities & Voices

Where working data scientists talk shop.

Want a personalised version?

This roadmap is the same one our platform uses internally, but the logged-in version lets you tick off topics as you complete them, track a personalised First 90 Days plan, and see your AI-durability score against this role. All free.

Open the interactive roadmap →
Curated by WhatTNext Ai · methodology · all roadmaps · last updated 2026-07-29