Oligarchy Project

Image Preprocessing

• Binarise images
• Crop images into single columned text
• Correct Skew

Extract Text

• OCR with AWS Textract (yielded better results than Tesseract).
• Correct OCR mistakes with SpellChecker
• Split text into sections based on each person's name
• Categorise information based on keywords and patterns

Build Network

• Group similar locations or institutions
• Map connections between nodes based on number of shared affiliations
• Draw network using Gephi