Piecing the puzzle - Self-publishing Queryable Research Data on the Web

Publishing research on the Web accompanied by machine-readable data is one of the aims of Linked Research. Merely embedding metadata as RDFa in HTML research articles, however, does not solve the problems of accessing and querying that data. Hence, I created a simple ETL pipeline to extract and enrich Linked Data from my personal website, publishing the result in a queryable way through Triple Pattern Fragments. The pipeline is open source, uses existing ontologies, and can be adapted to other websites. In this article, I discuss this pipeline, the resulting data, and its possibilities for query evaluation on the Web. More than 35,000 RDF triples of my data are queryable, even with federated SPARQL queries because of links to external datasets. This proves that researchers do not need to depend on centralized repositories for readily accessible (meta-)data, but instead can—and should—take matters into their own hands. INTRODUCTION The World Wide Web continues to shape many domains, and not in the least research. On the one hand, the Web beautifully fulfills its role as a distribution channel of scientific knowledge, for which it was originally invented. This spurs interesting dialogues concerning Open Access [1] and even piracy [2] of research articles. On the other hand, the advent of social networking creates new interaction opportunities for researchers, but also forces us to consider our online presence [3]. Various social networks dedicated to research have emerged: Mendeley, ResearchGate, Academia, ... They attract millions of researchers, and employ various tactics to keep us there. A major issue of these social research networks is their lack of mutual complementarity. None of them has become a clear winner in terms of adaption. At first sight, the resulting plurality seems a blessing for diversity, compared to the monoculture of Facebook for social networking in general. Yet whereas other generic social networks such as Twitter and LinkedIn serve complementary professional purposes compared to Facebook, social research networks share nearly identical goals. As an example, a researcher could announce a newly accepted paper on Twitter, discuss its review process on Facebook, and share a photograph of an award on LinkedIn. In contrast, one would typically not exclusively list a specific publication on Mendeley and another on Academia, as neither publication list would be complete. In practice, this results in constant bookkeeping for researchers who want each of their profiles to correctly represent them—a necessity if such profiles are implicitly or explicitly treated as performance indicators [4]. Deliberate absence on any of these networks is not a viable option, as parts of one’s publication metadata might be automatically harvested or entered by co-authors, leaving an automatically generated but incomplete profile. Furthermore, the quality of such non-curated metadata records can be questionable. As a result, researchers who do not actively maintain their online research profiles risk ending up with incomplete and inaccurate publication lists on those networks. Such misrepresentation can be significantly worse than not being present at all—but given the public nature of publication metadata, complete absence is not an enforceable choice. Online representation is not limited to social networks: scientific publishers also make metadata available about their journals and books. For instance, Springer Nature recently released SciGraph, a Linked Open Data platform that includes scholarly metadata. Accuracy is less of an issue in such cases, as data comes directly from the source. However, quality and usability are still influenced by the way data is modeled and whether or how identifiers are disambiguated. Completeness is not guaranteed, given that authors typically target multiple publishers. Therefore, even such authoritative sources do not provide individual researchers with a correct profile. In the spirit of decentralized social networking [5] and Linked Data [6], several researchers instead started publishing their own data and metadata. I am one of them, since I believe in practicing what we preach [7] as Linked Data advocates, and because I want my own website to act as the main authority for my data. After all, I can spend more effort on the completeness and accuracy of my publication metadata than most other platforms could reasonably do for me. In general, self-published data typically resides in separate RDF documents [8] (for which the FOAF vocabulary [9] is particularly popular [10]), or inside of HTML documents (using RDFa Lite [11] or similar formats). Despite the controllable quality of personally maintained research data and metadata in individual documents on the Web, they are not as visible, findable, and queryable as those of social research networks. I call a dataset interface “queryable” with respect to a given query when a consumer does not need to download the entire dataset in order to evaluate that query over it with full completeness. Unfortunately, hosting advanced search interfaces on a personal website quickly becomes complex and expensive. To mitigate this, I have implemented a simple Extract/Transform/Load (ETL) pipeline on top of my personal website, which extracts, enriches, and publishes my Linked Data in a queryable way through a Triple Pattern Fragments [12] interface. The resulting data can be browsed and queried live on the Web, with higher quality and flexibility than on my other online profiles, and at only a limited cost for me as data publisher. This article describes my use case, which resembles that of many other researchers. I detail the design and implementation of the ETL pipeline, and report on its results. At the end, I list open questions regarding self-publication, before concluding with a reflection on the opportunities for the broader research community. Piecing the puzzle Self-publishing queryable research data on the Web