This wiki service has now been shut down and archived
Reading Experience Database (RED) Author
From ESIWiki
Contents |
Name
Reading Experience Database (RED) Author
Owner
Mathieu D’Aquin [m.daquin@open.ac.uk]
Linked Open Data Source
http://data.open.ac.uk/ (data derived from original data source http://www.open.ac.uk/Arts/RED/) and DBPedia (http://dbpedia.org)
Background
The Reading Experience Database (RED) is “an open-access database housed at The Open University containing over 30,000 easily searchable records documenting the history of reading in Britain from 1450 to 1945” (http://www.open.ac.uk/Arts/RED/). It includes biographical information about people, what they have read, where and when they read it, and bibliographic details of the things they have read (including the author of the work).
As part of the Lucero Project (http://lucero-project.info), which is creating Linked Data at the Open University, the RED data has been expressed as Linked Data, and a demonstration application has been created – the ‘RED Author’ application.
Use case Scenario
The RED Author application is intended to demonstrate how links to external datasets can be useful from the point of view of research data. It is based on a draft dataset extracted from the Reading experience dataset.
RED Author shows a person (author or reader), with information from the database (reading experiences and books written) and from DBpedia (abstract, categories, influences). Clicking a category add to the page links to other people in the reading experience database that are also members of this category.
More details and a link to the demonstration application are available at http://lucero-project.info/lb/2011/03/connecting-the-reading-experience-database-to-the-web-of-data/.
Problems and Limitations
The nature of DBpedia means that the categories available may differ significantly from one author/person to another. This is especially true considering that only well known authors will be in adequately represented in DBpedia, while less well known authors, and many readers with entries in RED will not have entries in DBpedia.
At the moment, the application is intended as a demonstrator of the potential of connecting data collected and managed as part of research projects to external data. As such the integration (linking, querying and presentation) of external data (i.e., DBPedia) is realised in an ad-hoc manner. However, ideally, integrating other relevant sources of data (e.g., to relative the geographical aspects of reading or to historical events at the times of reading experiences) should be achieved without major development work.
Requirements for Provenance
Using DBpedia categories open up new avenues of research for RED, by exploiting the ‘long tail’ of information about people featured in RED. The quality of data required when querying the data for academic publication is high, and researchers will wish to be able to check the sources of the data easily before drawing firm conclusions based on the data. The issue is less trivial than it might first appear, as a research process might include complex queries (or series of queries) which results would draw both from data collected by academics from verifiable sources and from other external Web datasets (DBPedia in our case). It is therefore necessary to be able to trace and present to researchers the ‘evidence’ for such results, as well as their origin. The expectation is that, in the end, researchers would primary use external Web data as a way to identify possible research questions, areas and connections to explore, to follow up with more rigorous data analysis processes to support their initial conclusions.
The same aspect is also seen from the inverse perspective, i.e. the one of data exposure. Indeed, a large part of the work of researchers (especially in Arts and Humanities) is to collect data from a large variety of sources (especially archives), in a rigorous and traceable process. In order to make such datasets as useful and exploitable as possible, information about the data collection process should also be expose together with the data, as evidence of their trustworthiness. Indeed, as sometimes expressed by researchers we worked with, “we don’t want our data to be mixed up with Wikipedia”.