Data Science: Working with Data Course

Lesson 1 • Tools and Environment Setup

Introduces the standard data science toolchain and configures a reproducible working environment. Ensures all students operate from a consistent technical baseline.

Lesson 2 • The Data Science Lifecycle

Traces a project from business problem to deployed insight. Provides a repeatable mental model for structuring any data science engagement.

Lesson 3 • Types of Data and Problems

Categorises structured, unstructured, and semi-structured data and maps each to problem types. Guides appropriate method selection from the start.

Lesson 1 • Data Sources and Formats

Surveys common data origins and file formats encountered in practice. Prepares students to handle any incoming data format confidently.

Lesson 2 • Consuming APIs and Web Data

Covers REST API calls, authentication, and HTML scraping for data collection. Expands the range of data sources students can access programmatically.

Lesson 3 • Data Storage Strategies

Compares flat files, relational databases, and columnar stores for different workloads. Students select appropriate storage based on volume and query patterns.

Lesson 4 • Querying Relational Databases

Teaches SQL fundamentals for extracting and filtering data from relational stores. Directly enables data retrieval tasks in every subsequent chapter.

Lesson 1 • Data Type Conversion and Formatting

Addresses parsing dates, encoding categories, and standardising text fields. Ensures downstream tools receive correctly typed inputs.

Lesson 2 • Reshaping and Merging Datasets

Teaches pivoting, melting, and joining multiple tables into a unified analytical dataset. Directly prepares data for exploratory analysis in the next chapter.

Lesson 3 • Detecting and Treating Outliers

Covers statistical and visual methods for identifying anomalous values and deciding on treatment. Protects model performance from distorted inputs.

Lesson 4 • Handling Missing Data

Explains mechanisms of missingness and appropriate imputation or removal strategies. Prevents bias introduced by naive missing-value treatment.

Lesson 1 • Univariate Visual Analysis

Teaches histograms, box plots, and density plots for single-variable distributions. Builds visual literacy applied throughout the course.

Lesson 2 • Bivariate and Multivariate Analysis

Explores relationships between two or more variables using scatter plots, heatmaps, and pair plots. Surfaces correlations and interactions relevant to modelling.

Lesson 3 • Categorical Data Exploration

Analyses frequency distributions and cross-tabulations for categorical variables. Completes the EDA toolkit for mixed data types.

Lesson 4 • Descriptive Statistics

Covers central tendency, dispersion, and distributional shape for numeric variables. Provides the quantitative foundation for all subsequent visual analysis.

Lesson 1 • Scaling and Normalisation

Applies min-max scaling, standardisation, and robust scaling to numeric features. Ensures distance-based and gradient-based models receive comparable inputs.

Lesson 2 • Encoding Categorical Variables

Covers one-hot, ordinal, and target encoding for converting categories into numeric inputs. Prevents encoding-induced bias in downstream models.

Lesson 3 • Dimensionality Reduction

Introduces PCA and other reduction techniques to compress high-dimensional feature spaces. Reduces overfitting risk and computational cost before modelling.

Lesson 4 • Creating New Features

Derives interaction terms, polynomial features, and domain-specific aggregates. Expands the feature space with signals not present in raw columns.

Lesson 1 • Cross-Validation and Overfitting

Applies k-fold and stratified cross-validation to detect overfitting and estimate generalisation. Produces reliable performance estimates before final model selection.

Lesson 2 • Classification Algorithms

Introduces logistic regression, decision trees, random forests, and support vector machines for categorical targets. Extends the supervised workflow to discrete outcomes.

Lesson 3 • Hyperparameter Tuning

Uses grid search, random search, and Bayesian optimisation to find optimal model configurations. Maximises model performance within computational constraints.

Lesson 4 • Model Evaluation Metrics

Defines accuracy, precision, recall, F1, ROC-AUC, RMSE, and R-squared for assessing model quality. Enables objective comparison across algorithms and configurations.

Lesson 1 • Evaluating Cluster Solutions

Applies internal and external validation metrics to assess and compare cluster solutions. Enables defensible cluster selection without ground-truth labels.

Lesson 2 • Density and Hierarchical Clustering

Applies DBSCAN for noise-tolerant clustering and hierarchical methods for dendrogram-based segmentation. Handles non-spherical and nested cluster structures.

Lesson 3 • Clustering Fundamentals

Explains the unsupervised learning paradigm and the role of distance metrics in grouping. Provides the conceptual grounding for all clustering algorithms covered.

Lesson 4 • Interpreting and Using Cluster Outputs

Profiles cluster characteristics and integrates cluster labels as features for supervised models. Connects unsupervised findings to business decisions and downstream pipelines.

Lesson 1 • Model Governance and Documentation

Establishes model cards, audit trails, and access controls for responsible production use. Satisfies organisational and regulatory accountability requirements.

Lesson 2 • Monitoring Model Performance

Tracks prediction quality, data drift, and system health metrics in production. Detects degradation before it impacts business outcomes.

Lesson 3 • Containerisation and Deployment Pipelines

Packages model services in containers and automates deployment through CI/CD pipelines. Reduces manual deployment errors and accelerates iteration cycles.

Lesson 4 • Preparing Models for Production

Covers serialisation, dependency management, and reproducibility requirements for production handoff. Ensures models behave identically outside the development environment.