Data Science

6 subjects · 13 chapters · 52 topics

1. 01

   The Data Science Workflow

   1.1 How a project runs

   • Problem framing and success metrics
   • CRISP-DM and modern variants
   • Roles in a data team
   • Reproducibility from the start

   1.2 Tooling

   • Python, Jupyter and virtual environments
   • Git for data projects
   • NumPy and Pandas
   • Environment and dependency management
2. 02

   Working with Data

   2.1 Acquisition

   • Files, databases and APIs
   • Web scraping responsibly
   • Data contracts and schemas
   • Storing raw data safely

   2.2 Cleaning and preparation

   • Missing data strategies
   • Outliers and winsorising
   • Encoding categorical variables
   • Scaling and normalisation

   2.3 Feature engineering

   • Creating features from dates and text
   • Interaction and polynomial features
   • Feature selection
   • Leakage and how to avoid it
3. 03

   Statistics and Inference

   3.1 Descriptive and inferential

   • Distributions and the central limit theorem
   • Sampling and standard error
   • Hypothesis tests
   • Confidence intervals

   3.2 Experimentation

   • A/B test design
   • Power and sample size
   • Multiple comparisons
   • Reading results honestly
4. 04

   Modelling

   4.1 Supervised learning

   • Linear and logistic regression
   • Trees, random forests and gradient boosting
   • Train/validation/test discipline
   • Cross-validation

   4.2 Evaluation

   • Accuracy, precision, recall and F1
   • ROC and precision-recall curves
   • Regression metrics
   • Choosing a metric that matches the business

   4.3 Unsupervised learning

   • Clustering
   • Dimensionality reduction
   • Association rules
   • Interpreting unsupervised output
5. 05

   Communication and Delivery

   5.1 Visualisation

   • Matplotlib and Seaborn
   • Plotly for interactivity
   • Dashboard design
   • Charts that do not mislead

   5.2 Putting a model to work

   • Serialising a model
   • A simple prediction API
   • Batch versus real-time scoring
   • Monitoring drift
6. 06

   Capstone

   6.1 End-to-end project

   • Choosing a dataset and a question
   • Building and validating the model
   • Writing the report
   • Presenting the findings