![What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata
Top Metadata Errors
10
25% of the datasets access information are not clean [unavailability mainly]
68% of the resources access data can be fixed automatically
31.27% of the resources were not reachable
63.17% of the resources don’t have valid resource_type
Creating an aggregate report for resources > format:title
62.16% of the datasets have defined SPARQL endpoints using the api/sparql resource format
92.27% provided RDF example links
56.3% provided direct links to downloadable RDF dumps](https://image.slidesharecdn.com/sup-151204011748-lva1-app6892/85/What-s-up-LOD-Cloud-Observing-the-state-of-Linked-Open-Data-Cloud-Metadata-10-320.jpg)
What's up LOD Cloud - Observing the state of Linked Open Data Cloud Metadata
- 1. What's up LOD Cloud Observing The State of Linked Open Data Cloud Metadata Ahmad Assaf, Raphaël Troncy And Aline Senart LDQ15 – 2nd Workshop on Linked Data Quality 1st June 2015 @ahmadaassaf
- 2. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata LOD Cloud 2 LOD Cloud is a mine of data The heterogeneity of sources reflect directly on the data quality Finding useful dataset without prior knowledge is increasingly difficult Demonstrate the of LOD Cloud metadata by running Roomba on the accessible LOD Cloud through datahub.io
- 3. Dataset Metadata General information e.g. title, description Ownership information e.g. author, maintainer_email Provenance information e.g. creation_date, version Access information e.g. license_title, license_id Metadata is structured information that describes, explains, locates or otherwise makes it easier to retrieve use or manage information resources Data Portals are a curated collection of datasets metadata providing a set discovery and integration services We divided the metadata information into four main types 3What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata
- 4. Metadata Information Groups 4 Organization Clustering or curation solely based on associations with specific administration parties Resource Actual raw data that can be downloaded or accessed directly e.g. JSON, CSV, SPARQL endpoint Tag Descriptive knowledge about the dataset contents and structure. This can range from simple textual tags to semantically rich controlled terms Group Organizational units that share common semantics. They can be seen as a cluster or curation based on shared themes/categories What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata
- 5. Roomba - An Extensible Framework to Validate and Build Dataset Profiles Roomba addresses the challenges of automatic validation and generation of descriptive dataset profiles https://github.com/ahmadassaf/opendata-checker/ 5What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata
- 6. 6
- 7. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Metadata Errors for Information Groups 7 41% of ownership information is missing or undefined Resources have the poorest metadata health across information groups 64% general metadata 100% access metadata 80% provenance metadata
- 8. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Top Metadata Errors 8 19% of these errors can be fixed automatically 33.33% can be fixed automatically by tools plugged into the data publishing workflow
- 9. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Top Metadata Errors 9 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 %ofresources Information Field
- 10. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Top Metadata Errors 10 25% of the datasets access information are not clean [unavailability mainly] 68% of the resources access data can be fixed automatically 31.27% of the resources were not reachable 63.17% of the resources don’t have valid resource_type Creating an aggregate report for resources > format:title 62.16% of the datasets have defined SPARQL endpoints using the api/sparql resource format 92.27% provided RDF example links 56.3% provided direct links to downloadable RDF dumps
- 11. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Metadata File Types Errors 11
- 12. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Metadata Errors 12 Noisiest part of the access metadata is license information 16.6% of the datasets don’t have license_title and license_id 54.44% datasets don’t have license_url 51.35% don’t have maintainer 55.21% of the datasets are missing maintainer_email | 15.06% author_email 80% of the provenance information is missing or undefined The only manual field in provenance information “version” is missing from 60.23% of the datasets
- 13. What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata Enriched Profiles 1.87% of the resources have incorrect mimetype defined 4.82% of the resources have incorrect size values 47.49% of the datasets license information have been normalized via the manual license mapping file ✝ 13 ✝ https://github.com/ahmadassaf/opendata-checker/blob/master/util/licenseMappings.json
- 14. 14 Questions? Ahmad Assaf http://ahmadassaf.com @ahmadaassaf http://github.com/ahmadassaf What's up LOD Cloud - Observing The State of Linked Open Data Cloud Metadata