Metadata Crosswalk Visualizer

A visual, standards‑aware tool for understanding how bibliographic metadata maps — and degrades — across major library metadata frameworks.

The Metadata Crosswalk Visualizer shows how core bibliographic concepts translate between UNIMARC, MARC 21, Dublin Core, RDA, and BIBFRAME 2.0, highlighting where data is preserved, partially preserved, or irretrievably lost during conversion.

It is designed for metadata professionals who need more than generic crosswalk tables: it exposes structure, semantics, and real‑world system consequences.

---

✨ Key Features

• Concept‑based navigation
  Explore mappings by bibliographic concept (Title, Creator, Subject, Identifier, etc.), not just by field number.

• Five‑standard comparison
  Side‑by‑side mappings across:

  • UNIMARC
  • MARC 21
  • Dublin Core (Simple & Qualified)
  • RDA
  • BIBFRAME 2.0

• Match quality indicators
  Each mapping is classified as:

  • Exact – semantic structure preserved
  • Partial – some loss or flattening
  • Context‑dependent – meaning depends on implementation
  • Lossy – significant semantic loss
  • No equivalent

• System‑level notes
  Practical observations from real systems, including Koha and Libris XL, documenting where migrations and exports fail in practice.

• Real‑world problem scenarios
  Concrete examples of how crosswalk limitations affect discovery, authority control, multilingual data, and classification integrity.

• Interactive crosswalk mode
  Compare a source standard against a target standard to immediately see where and how data loss occurs.

• Full‑text search
  Search by field number (e.g. 245, 650), subfield ($a), element name, or keyword.

---

🎯 Who This Is For

• Metadata librarians and catalogers
• Library systems and migration teams
• LIS educators and students
• Developers working with library metadata
• Anyone designing or evaluating metadata crosswalks

This tool assumes familiarity with bibliographic standards and is intentionally explicit about technical detail.

---

🧭 Design Principles

• Concepts over fields
  Users think in Title, Creator, Subject — not just 245 or dc:title.

• Semantic honesty
  Not all mappings are equal. The tool makes data loss visible instead of hiding it.

• Standards‑first
  Mappings are grounded in official documentation (LC, IFLA, DCMI, RDA Toolkit), cross‑checked against professional practice.

• System reality matters
  What works “on paper” often fails in ILS, LSP, or aggregation environments.

---

🛠️ Technology

• React (single‑page application)
• No backend
• No external dependencies beyond React
• Designed to be embedded in a larger React app or deployed as a static site

The main implementation lives in a single component file: