What is LOD & Open Data?
LOD (Linked Open Data) and open data, told as a story about a warehouse that opens its door and links its tags.
Nobody outside knows what's in this warehouse.The door is locked, and there's no label either.
The data is published, but no program can read it.When data formats don't match, reusing the data is hard.
Open data/LOD opens the warehouse door, tags data with a standard ID, and links it for anyone.Anyone can come and reuse the data.
Condensing the 5-star model, open data gets easier to reuse through four steps, from photo/PDF up to linked data.The last step is exactly LOD (Linked Open Data).
Publishing data openly doesn't guarantee its quality.Open formats make data easy to find, but don't guarantee the content is accurate.
Open data/LOD opens the warehouse door to everyone, tags boxes with a standard ID (URI), and links them to other warehouses.
Open data is data published so anyone can freely access, reuse, and redistribute it. LOD (Linked Open Data), the top tier of Tim Berners-Lee's 5-star model, assigns URIs to data and links it to other data via RDF, forming a connected web of open data. Publishing it openly is a separate matter from its quality or freshness.
When grown-ups say it
- Open Data
- A warehouse opened for anyone to use. Real openness needs a clear license too.
- Linked Open Data
- Warehouses tagged and linked to each other. The highest tier of the 5-star model.
- 5-Star Open Data
- Five steps that make open data easier to reuse. More stars, easier reuse.
- URI
- The standard tag attached to every box. Lets different programs point to the same box.
- RDF
- The rule for the string linking box to box. Written as subject–predicate–object.
- NoSQL Modeling
- A warehouse reshaping shelves to fit each box. Graph-type NoSQL fits linked data well. → the free-shelf warehouse
- Normalization
- Writing shared info on one shelf once. Open data is easier to manage with less duplication too. → the rule of writing it once
- Personal Data / Privacy
- Information that must be masked before publishing. Masking personal data is covered in the AI field, with AI as the example. → AI privacy