Phase 28: Practical Data Science

This folder is the transition from studying concepts to performing applied work under interview-style and project-style constraints. It is broad by design, so the main job of this README is to keep the breadth from turning into noise.

What This Phase Is For

  • Consolidating ML, statistics, SQL, forecasting, and applied deep learning

  • Practicing end-to-end workflows instead of isolated techniques

  • Preparing for interviews, take-homes, and portfolio projects

Subtracks

  • python-data-science/: data cleaning, EDA, feature work

  • machine-learning/: pipelines, validation, ensembles, interpretability

  • statistics-mlops/: experimentation, metrics, monitoring thinking

  • sql-data-engineering/: data extraction and transformation habits

  • time-series-forecasting/: forecasting and anomaly detection

  • computer-vision/: practical CV workflows

  • deep-learning-nlp/: applied NLP and deep learning practice

  • recommender-causal/: recommendation and causal framing

  • solutions/: answer keys or worked examples where available

How To Use This Folder Well

  • Pick a role first, then limit yourself to the most relevant subtracks.

  • Prefer one complete project over sampling ten notebooks shallowly.

  • Write up your assumptions, metrics, and trade-offs as if someone else will review your work.

Capstone Ideas

  • Churn prediction with a full validation and monitoring plan

  • Forecasting dashboard with anomaly alerts

  • Recommendation prototype with offline evaluation

  • SQL-to-model pipeline with a short stakeholder-facing writeup