Purpose The purpose of this paper is to employ the case of Organization for Economic Cooperation and Development (OECD) data repositories to examine the potential of blockchain technology in the context of addressing basic contemporary societal concerns, such as transparency, accountability and trust in the policymaking process. Current approaches to sharing data employ standardized metadata, in which the provider of the service is assumed to be a trusted party. However, derived data, analytic processes or links from policies, are in many cases not shared in the same form, thus breaking the provenance trace and making the repetition of analysis conducted in the past difficult. Similarly, it becomes tricky to test whether certain conditions justifying policies implemented still apply. A higher level of reuse would require a decentralized approach to sharing both data and analytic scripts and software. This could be supported by a combination of blockchain and decentralized file system technology. Design/methodology/approach The findings presented in this paper have been derived from an analysis of a case study, i.e., analytics using data made available by the OECD. The set of data the OECD provides is vast and is used broadly. The argument is structured as follows. First, current issues and topics shaping the debate on blockchain are outlined. Then, a redefinition of the main artifacts on which some simple or convoluted analytic results are based is revised for some concrete purposes. The requirements on provenance, trust and repeatability are discussed with regards to the architecture proposed, and a proof of concept using smart contracts is used for reasoning on relevant scenarios. Findings A combination of decentralized file systems and an open blockchain such as Ethereum supporting smart contracts can ascertain that the set of artifacts used for the analytics is shared. This enables the sequence underlying the successive stages of research and/or policymaking to be preserved. This suggests that, in turn, and ex post, it becomes possible to test whether evidence supporting certain findings and/or policy decisions still hold. Moreover, unlike traditional databases, blockchain technology makes it possible that immutable records can be stored. This means that the artifacts can be used for further exploitation or repetition of results. In practical terms, the use of blockchain technology creates the opportunity to enhance the evidence-based approach to policy design and policy recommendations that the OECD fosters. That is, it might enable the stakeholders not only to use the data available in the OECD repositories but also to assess corrections to a given policy strategy or modify its scope. Research limitations/implications Blockchains and related technologies are still maturing, and several questions related to their use and potential remain underexplored. Several issues require particular consideration in future research, including anonymity, scalability and stability of the data repository. This research took as example OECD data repositories, precisely to make the point that more research and more dialogue between the research and policymaking community is needed to embrace the challenges and opportunities blockchain technology generates. Several questions that this research prompts have not been addressed. For instance, the question of how the sharing economy concept for the specifics of the case could be employed in the context of blockchain has not been dealt with. Practical implications The practical implications of the research presented here can be summarized in two ways. On the one hand, by suggesting how a combination of decentralized file systems and an open blockchain, such as Ethereum supporting smart contracts, can ascertain that artifacts are shared, this paper paves the way toward a discussion on how to make this approach and solution reality. The approach and architecture proposed in this paper would provide a way to increase the scope of the reuse of statistical data and results and thus would improve the effectiveness of decision making as well as the transparency of the evidence supporting policy. Social implications Decentralizing analytic artifacts will add to existing open data practices an additional layer of benefits for different actors, including but not limited to policymakers, journalists, analysts and/or researchers without the need to establish centrally managed institutions. Moreover, due to the degree of decentralization and absence of a single-entry point, the vulnerability of data repositories to cyberthreats might be reduced. Simultaneously, by ensuring that artifacts derived from data based in those distributed depositories are made immutable therein, full reproducibility of conclusions concerning the data is possible. In the field of data-driven policymaking processes, it might allow policymakers to devise more accurate ways of addressing pressing issues and challenges. Originality/value This paper offers the first blueprint of a form of sharing that complements open data practices with the decentralized approach of blockchain and decentralized file systems. The case of OECD data repositories is used to highlight that while data storing is important, the real added value of blockchain technology rests in the possible change on how we use the data and data sets in the repositories. It would eventually enable a more transparent and actionable approach to linking policy up with the supporting evidence. From a different angle, throughout the paper the case is made that rather than simply data, artifacts from conducted analyses should be made persistent in a blockchain. What is at stake is the full reproducibility of conclusions based on a given set of data, coupled with the possibility of ex post testing the validity of the assumptions and evidence underlying those conclusions.
Purpose Nowadays, database management system has been applied in library management, and a great number of data about readers’ visiting history to resources have been accumulated by libraries. A lot of important information is concealed behind such data. The purpose of this paper is to use a typical data mining (DM) technology named an association rule mining model to find out borrowing rules of readers according to their borrowing records, and to recommend other booklists for them in a personalized way, so as to increase utilization rate of data resources at library. Design/methodology/approach Association rule mining algorithm is applied to find out borrowing rules of readers according to their borrowing records, and to recommend other booklists for them in a personalized way, so as to increase utilization rate of data resources at library. Findings Through an analysis on record of book borrowing by readers, library manager can recommend books that may be interested by a reader based on historical borrowing records or current book-borrowing records of the reader. Research limitations/implications If many different categories of book-borrowing problems are involved, it will result in large length of encoding as well as giant searching space. Therefore, future research work may be considered in the following aspects: introduce clustering method; and apply association rule mining method to procurement of book resources and layout of books. Practical implications The paper provides a helpful inspiration for Big Data mining and software development, which will improve their efficiency and insight on users’ behavior and psychology. Social implications The paper proposes a framework to help users understand others’ behavior, which will aid them better take part in group and community with more contribution and delightedness. Originality/value DM technology has been used to discover information concealed behind Big Data in library; the library personalized recommendation problem has been analyzed and formulated deeply; and a method of improved association rules combined with artificial bee colony algorithm has been presented.
Purpose - The purpose of this paper is to present the knowledge structure based on the articles published in Library Hi Tech. The research hotspots are expected to be revealed through the keyword co-occurrence and social network analysis. Design/methodology/approach - Data sets based on publications from Library Hi Tech covering the time period from 2006 to 2017 were extracted from Web of Science and developed as testbeds for evaluation of the CiteSpace system. Highly cited keywords were analyzed by CiteSpace which supports visual exploration with knowledge discovery in bibliographic databases. Findings - The findings suggested that the percentage of publications in the USA, Germany, China, and Canada are high. Further, the most popular keywords identified in Library Hi Tech were: "service," "technology," "digital library," "university library," and "academic library." Finally, four research issues were identified based on the most-cited articles in Library Hi Tech. Originality/value - While keyword plays an important role in scientific research, limited studies paid attention to the keyword analysis in librarian research. The contribution of this study is to systematically explore the knowledge structure constructed by the keywords in Library Hi Tech.