Triply

GO FAIR

Wouter Beek (wouter@triply.cc),
triply.cc, triplydb.com

Knowledge Graph

The Network Effect for Data

The value of a network is proportional to the square of the number of connected nodes.

Linked Data

lod-cloud.net

Data is already FAIR for GAFAM

“Inside the Alexa-Friendly World of Wikidata” Wired 2019-02-18 (link)

Can a FAIR data platform be implemented?

Can it be done with Linked Data technology?

How can we bring FAIR data to everybody else?

Large-scale Knowledge Graph Hosting

  • 2014 LOD Laundromat: 38B edges
  • 2015 LOD Lab: reproduce 3 SW papers at scale
  • 2016 LOD Search: text index
  • 2017 LOD-a-lot: 28B edges on a USB thumbdrive
  • 2018 sameAs.cc: identity closure of 35B edges

TriplyDB: Large-scale Knowledge Graph Hosting

triply.cc
“The Semantic Web has lacked an essential element which the WWW had from the start: the immediate gratification for information providers to see the results of their efforts on a screen.”
Tim Berners-Lee “Tabulator Redux” (2007)

Geospatial information

Kadaster Knowledge Graph: key registries, company registry, monument registry, energy labels, etc.

Statistical information

Life expectancy for the countries of the world over time (dataset: Gapminder).

Temporal information

Timeline of cars produced by Ford, Chevrolet, Porsche, and Toyota between 1903 and 2016 (dataset: DBpedia).

Linked Data cost/benefit

Cost

“Data are described with rich metadata.”

Benefit

Findable using popular search engines.

Linked Data cost/benefit

Cost

“(Meta)data are assigned a globally unique and eternally persistent identifier.”

Benefit

Zero-cost data integration.

Data Stories

A sequence of real-time generated data visualizations, connected through an overarching story.

Data Stories are…

  • reproducible
  • modifiable
  • standardized
  • explainable

Data Story Example: Churches in NL

Data Story Example: Maritime Sources

Data Stories → FAIR paper

  • HTML5 paper (RASH) + metadata
  • Store data together with the paper
  • Institutional embedding
  • Community acceptance

Summary

  • Bring down the cost of hosting FAIR data.
  • FAIR data must bring an immediate reward to the data publisher (e.g., constructive data search).
  • The (perceived) benefits of FAIR must exceed the (perceived) costs of FAIR.
  • Make it difficult to publish data that is not FAIR.
  • FAIR paper: use of FAIR data.

Thank you for your attention!

Wouter Beek (wouter@triply.cc)