Learn Python and data analysis
Get comfortable working with tabular data before training models.
- Python
- NumPy & pandas
- SQL
- Exploratory analysis
Your checkpoint
Clean a dataset and explain its important patterns and quality issues.
All roadmapsData Science & Machine Learning
Data Science & Machine Learning
Roadmap.
Build towards an ML portfolio with trained models, test results and demos you can take into internship interviews.
Build the maths you need
Connect statistical ideas to real questions. Learn to describe uncertainty rather than hiding it.
- Probability
- Statistics
- Sampling
- Statistical inference
Your checkpoint
Explain what a sample can and cannot tell you about a population.
Start with a baseline
Define a prediction task, choose a sensible metric and build a simple reference model.
- Regression
- Classification
- Train/test split
- Baseline models
Your checkpoint
Compare a trained model against a simple baseline on held-out data.
Evaluate and improve
Look at errors, leakage and overfitting. Choose improvements based on evidence from validation.
- Cross-validation
- Feature engineering
- Error analysis
- Model selection
Your checkpoint
Explain where your model fails and how your evaluation avoids leakage.
Explore different problems
Apply modelling to groups, sequences or changing data while checking each method’s assumptions.
- Clustering
- Forecasting
- Model interpretation
- Data limitations
Your checkpoint
Create a second project and explain why its evaluation differs from your first.
Make your model inspectable
Package your work in a reproducible notebook and a small demo, with clear caveats.
- Reproducibility
- Model demos
- Documentation
- Portfolio
Your checkpoint
Let someone try a prediction and read how well the model performed.
Learn with guidance
Build your portfolio
with NYC.
Follow this path with live classes, project work and mentor feedback in Data Science & Machine Learning. Build the portfolio and the understanding to explain it.
