Peer Review Database Cleanup — Proof of Concept

Technical prototype portfolio — demonstration environment

Run the bundled sample peer-review assignment export below, which has several real planted defects. No AI step — every row is checked against the real defined schema (required fields, real email and calendar-date validity), every row is compared against every other row sharing the same real reviewer-email-plus-manuscript-ID key for an exact or conflicting duplicate, and one representative row per distinct key is compared against every other for a real Levenshtein near-duplicate match on reviewer name. Nothing is silently removed: the exported, migration-ready CSV marks every row's real status, decided only from findings a reviewer has actually confirmed. Do not submit personal, confidential or client-owned information — this public demo only runs the bundled sample. Contact Chrisian to discuss a controlled evaluation.

Bundled sample records

sample-peer-review-records.csv — a short peer-review assignment export built with planted schema, duplicate, and near-duplicate defects to exercise this PoC's checks.

This PoC's bundled real schema (fixed, in this column order)

Reviewer Name — required, also checked for near-duplicate typos
Reviewer Email — required, email format, part of the duplicate-matching key
Manuscript ID — required, part of the duplicate-matching key
Assigned Date — required, date format
Status — required