All selected work
- Project
- Apr - May 2026
- Course project
Twenty Constitutions Digitalization
CPE232 Data Models
A document-engineering workflow that turns Thai constitutional archives into structured, reusable data.
Key numbers
- Thai constitutions
- 20
- source PDFs
- 38
- words processed
- ~61K
Highlights
- ExtractionOCR for scanned pages and direct text extraction for digital PDFs.
- NormalizationThai Unicode and Buddhist Era dates normalized, with document noise removed.
- StructureChapter, article, and section metadata for every document.
- OutputsValidated JSON and CSV for exploratory analysis and downstream NLP.
Evidence
Structured outputsSlide 18, JSON hierarchy and section-level CSV
Exploratory topic analysisSlide 28, keyword and embedding-based analysis
Processing pipelineFrom 38 source PDFs to JSON and CSV
