Ecole d'ingénieur et centre de recherche en Sciences du numérique

HDL - Towards a harmonized dataset model for open data portals

Assaf, Ahmad; Troncy, Raphaël; Senart, Aline

PROFILES 2015, 2nd International Workshop on Dataset Profiling & Federated Search for Linked Data, Main conference ESWC15, 31 May-4 June 2015, Portoroz, Slovenia

The Open Data movement triggered an unprecedented amount of data published in a wide range of domains. Governments and corporations around the world are encouraged to publish, share, use and integrate Open Data. There are many areas where one can see the added value of Open Data, from transparency and self-empowerment to improving efficiency, e ffectiveness and decision making. This growing amount of data requires rich metadata in order to reach its full potential. This metadata enables dataset discovery, understanding, integration and maintenance. Data portals, which are considered to be datasets' access points, o ffer metadata represented in di fferent and heterogenous models. In this paper, we fi rst conduct a unique and comprehensive survey of seven metadata models: CKAN, DKAT, Public Open Data, Socrata, VoID, DCAT and Next, we propose HDL, an harmonized dataset model based on this survey. We describe use cases that show the benefi ts of providing rich metadata to enable dataset discovery, search and spam detection.

Document Bibtex

Titre:HDL - Towards a harmonized dataset model for open data portals
Mots Clés:Dataset Metadata, Dataset Pro le, Dataset Model, Data Quality
Département:Data Science
Eurecom ref:4543
Copyright: CEUR
Bibtex: @inproceedings{EURECOM+4543, year = {2015}, title = {{HDL} - {T}owards a harmonized dataset model for open data portals}, author = {{A}ssaf, {A}hmad and {T}roncy, {R}apha{\"e}l and {S}enart, {A}line }, booktitle = {{PROFILES} 2015, 2nd {I}nternational {W}orkshop on {D}ataset {P}rofiling \& {F}ederated {S}earch for {L}inked {D}ata, {M}ain conference {ESWC}15, 31 {M}ay-4 {J}une 2015, {P}ortoroz, {S}lovenia}, address = {{P}ortoroz, {SLOV}{\'{E}}{NIE}}, month = {05}, url = {} }
Voir aussi: