Challenge
The Australian list of accredited visa sponsors was published only as scanned images, impossible to search or filter.
Selected work / Data engineering ยท Automation
Sponsor list OCR pipeline
Turned a 129 page scanned PDF into a clean, searchable dataset of 3,553 accredited sponsors.
The story
What I did
Converted pages to images with pdftoppm, extracted text with Tesseract OCR, then cleaned, deduplicated and standardised names with Python.
Result
3,553 unique sponsor names in a styled, filterable Excel workbook.
Tools
PythonTesseract OCRpdftoppmpandasExcel