Data Science using Python

Set up a reproducible data-science environment and master NumPy vectorisation, the linear algebra, calculus and probability intuition that every model depends on.

Pull data from files, databases, APIs and the web, then clean, merge, reshape and engineer it into tidy, analysis-ready datasets.

Profile data, expose patterns and outliers, and tell a clear visual story with Matplotlib, Seaborn and Plotly.

Reason under uncertainty: distributions, sampling, estimation, hypothesis testing, A/B testing and Bayesian thinking — with SciPy and statsmodels.

The complete supervised-learning workflow in scikit-learn: regression, classification, pipelines, cross-validation and honest model evaluation.

Ensembles and gradient boosting, clustering, dimensionality reduction, feature engineering and hyper-parameter tuning that win competitions.

Build neural networks from neuron to CNN with PyTorch and Keras — backpropagation, training tricks, transfer learning and computer vision.

Turn text into insight: preprocessing, embeddings, classification, and modern transformer models with spaCy and Hugging Face.

Decompose, test and forecast temporal data with ARIMA, exponential smoothing and machine-learning forecasters — and evaluate them properly.

Take a model to production: experiment tracking with MLflow, packaging with Docker, serving with FastAPI, and monitoring for drift.

Fairness, explainability, privacy and governance with the NIST AI RMF — plus a GitHub portfolio, case-study interviews and your data-scientist roadmap.