Data Scientist Roadmap: The 6 Stages

This is the roadmap we recommend for going from zero to employable data scientist. Work the stages in order — each one builds on the last. When you’re ready to pick courses for a stage, our data science courses guide maps options to every step below.

Stage 1 — Mathematics

• Linear algebra: matrices, vectors, core formulas
• Calculus: derivatives and gradients (what optimizers actually do)
• Probability & statistics: mean, median, mode, distributions, hypothesis testing

Stage 2 — Excel

• Data filters, charts and plots
• Pivot tables and transpose
• VBA macros for automation

Stage 3 — A programming language (Python or R)

• Core language: functions, dictionaries, tuples
• The data stack: NumPy, Pandas, Matplotlib
• Time complexity basics and working with databases (SQL)

Stage 4 — Data visualization

• Power BI or Tableau (pick one and go deep)
• Dashboarding and storytelling with data
• Big-data context: where Hadoop and distributed storage fit

Stage 5 — Machine learning

• scikit-learn: regression, classification, clustering
• TensorFlow 2 fundamentals
• Model evaluation and iteration — see our machine learning courses guide

Stage 6 — Deep learning

• Artificial neural networks (ANNs)
• Convolutional neural networks (CNNs) and recurrent neural networks (RNNs)
• TensorBoard and TensorFlow Hub for real projects — covered in our deep learning courses guide

You don’t need all six stages before applying for jobs — many analysts are hired after stage 4. But a portfolio project that runs the full pipeline, from raw data to a deployed model, is the strongest signal you can put in front of a hiring manager.