Learn why structured data is essential and how to move from unstructured documents to machine-readable formats. Explore CSV files, their limitations, and an introduction to markup languages for semantic annotation. This is the foundation for understanding databases and data modeling.
Structured vs Unstructured: A 517-page PDF of MSC projects is unstructured - you can't answer "How many supervisors?" without reading all 517 pages. Convert it to CSV and you can answer with cut -d',' -f1 file.csv | sort | uniq | wc -l in seconds.
CSV is powerful but limited: Perfect for flat lists you can process with Unix pipes. Breaks when you need relationships between entities (students ↔ modules ↔ coursework) or when data contains newlines/commas. Always escape with quotes!
Entities, Attributes, Relationships: The three pillars of data modeling. Entity = student, train, project. Attributes = student's exam score, train's speed. Relationships = student takes module, module has coursework. CSV can't model relationships well.
Markup adds semantics: Instead of plain text "Cat on train," markup says <story><subject>Cat</subject><location>London Euston</location><outcome>Happy</outcome></story>. Now machines can search, aggregate, predict without understanding English!