Manually converted 300+ scanned pages from the 1951 Census of India into 87 clean, structured Excel files — up to 40 data fields per village (population, literacy, land, industry).
Each village's data was split across 4 scanned pages, so I cross-referenced by Location Code Number and merged everything into one clean row per village.
Why manual, not OCR: faded 70+ year old print and ambiguous digits (e.g. '1,077.94' vs '1.077.94') make automated tools unreliable. I verified every ambiguous number against context and preserved blank cells accurately.
Available for: transcription of scanned/handwritten records, old ledgers, and archival documents where OCR fails — plus multi-page data reconciliation and large-volume data entry.