The problem
ESI's registrar's office kept decades of student records on paper only.
- Any lookup in the archive cost the staff hours of manual searching.
- No digital trace of the records existed.
- Names were spelled inconsistently across documents.
- Many student IDs were missing.
- Scan quality was low.
What we built
A full digitization process and archive management system for the 1977–1984 records, built by a team of 6 as our 2nd-year pluridisciplinary project.
1. Digitization pipeline (Python)
- Extracts data from scanned deliberation reports using image processing and LLM-assisted extraction.
- Verification scripts flag name inconsistencies and extraction errors.
2. Database (SQLite)
- Stores students, modules and results in a clean relational structure.
3. Desktop app (Flutter)
- Search students across the whole archive.
- Generate transcripts and certificates.
- View rankings and pull statistics.
- Runs fully offline, so the data stays on the office's machine.
Results
- 72 documents digitized.
- ~2,177 student transcripts, searchable in seconds instead of hours.
- 100+ modules recorded.
screenshots
