Clean Mixed Data from Scans

I have a collection of scanned documents that hold a mix of text and numerical information. I need every page accurately transcribed, validated, and cleaned so I can move straight into Data analysis without worrying about stray characters, inconsistent units, or misplaced figures.

Your task involves:
• Manually (or with reliable OCR) extracting the mixed data from each scan
• Cross-checking entries for completeness and correctness
• Normalising dates, numbers, and naming conventions so the entire file is consistent
• Flagging any ambiguities or unreadable sections for my review

Deliverables I expect:
• One fully cleaned, analysis-ready file labelled “Data analysis”
• A brief change log outlining key cleaning decisions or assumptions

Accuracy is more important than speed, but I would like incremental updates so we can tackle any issues early. If you have experience turning scanned paperwork into structured Excel, CSV, or database tables, that background will be useful here, though the final package only needs to satisfy the “Data analysis” requirement described above.