Unleashing the Potential: How Automated Metadata Creation Can Improve Your Data Quality and Accessibility


Elena Robu | Astun Technology

What is Metadata anyway? 🤔

unlabelled cans

Photo by YesMore Content on Unsplash


In simple terms, metadata is data about data.

What is Q-FAIR all about? 🎡

Quality

- is just a hyphen

Findable

Accessible

Interoperable

Reusable

Background

Why are we doing this? 🤷

Background

How do we do this now? 🕰️

Metadata Crawler open source plugin for Talend Open Studio (aka Crawler)

What are these fields? 📜

  • Abstract
  • Alternative Title
  • Bounding box
  • Constraints (usage and licensing)
  • Contact Details
  • Date
  • Geographic Extent (keyword)
  • INSPIRE Keywords (controlled)
  • Keywords (free text or controlled)
  • Lineage
  • Spatial Reference System
  • Resolution
  • Temporal Extent
  • Title
  • Topic Category (controlled)
  • Update Frequency

What are these fields? 📜

  • Abstract
  • Alternative Title
  • Bounding box
  • Constraints (usage and licensing)
  • Contact Details
  • Date
  • Geographic Extent (keyword)
  • INSPIRE Keywords (controlled)
  • Keywords (free text or controlled)
  • Lineage
  • Spatial Reference System
  • Resolution
  • Temporal Extent
  • Title
  • Topic Category (controlled)
  • Update Frequency

What are these fields? 📜

  • Abstract
  • Alternative Title
  • Constraints (usage and licensing)
  • Contact Details
  • Geographic Extent (keyword)
  • INSPIRE Keywords (controlled)
  • Keywords (free text or controlled)
  • Lineage
  • Resolution
  • Temporal Extent
  • Topic Category (controlled)
  • Update Frequency

What are these fields? 📜

  • Abstract
  • Alternative Title
  • Geographic Extent (keyword)
  • INSPIRE Keywords (controlled)
  • Keywords (free text or controlled)
  • Topic Category (controlled)

How do we automate these? 🧑‍💻

"industrial strength open source NLP in Python"

Geographic Keywords 🌍

NER model trained on OS Open Names

Pattern recognition model (postcodes)

Non-Geographic Keywords 🏷️

INSPIRE Keywords and Topic Categories 🔖

Text Classification

Everything Else

How will this work IRL? 🤔

Questions? 🙋

Photo by Hans-Jurgen Mager on Unsplash

Thank you! 😊

astuntechnology.com/contact-us

Experts in Place