Exploring Coverage and Distribution of Scholarly Identifiers on the Web
3 likes1,412 views
This document summarizes a study exploring the coverage and distribution of scholarly identifiers on the web. The study analyzed over 95,000 scholarly papers from arXiv, finding that DOIs had higher coverage than arXiv IDs. Papers with DOIs were more findable on Mendeley than those with only arXiv IDs. The combination of arXiv ID and DOI was also very common on Mendeley. Coverage and distribution of identifiers varied significantly between disciplines. Future work could expand the study to more fields and data sources to better understand differences.
![16
Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
Distribution of Disciplines on Mendeley (March 2011)
[Kraker et al.
2012]](https://image.slidesharecdn.com/6exploringscholarlyidentifierskraker-150520174758-lva1-app6892/85/Exploring-Coverage-and-Distribution-of-Scholarly-Identifiers-on-the-Web-16-320.jpg)
Exploring Coverage and Distribution of Scholarly Identifiers on the Web
- 1. 1 b b www.know-center.at Exploring Coverage and Distribution of Scholarly Identifiers on the Web 14th International Symposium of Information Science Zadar, 21 May 2015 Peter Kraker, Asura Enkhbayar & Elisabeth Lex
- 2. 2 Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics The Emerging Eco-system of Scholarly Services on the Web
- 3. 3 Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics The Emerging Eco-system of Scholarly Services on the Web
- 4. 4 Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 5. 5 Research Questions • How are scholarly identifiers distributed in crowd- sourced systems on the Web? • Does the provision of different identifiers have an influence on findability of scientific publications in other bibliographic and bibliometric sources? • Who are the top providers of identifiers? • Which identifier combinations are the most common? Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 6. 6 Indentifiers on the Scholarly Web Article level Publication levelAuthor level Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 7. 7 Data & Method Data Sources: arXiv, CrossRef and Mendeley arXiv discipline of quantitative biology (q-bio), 1992-2014 (n=14,195 metadata records) Data collection: 17/11/2014 arXiv discipline of astrophysics (astro-ph), 2009-2014 (n=81,814 metdata records) Data collection: 04/03/2015 Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics arxiv-ID, DOI DOI • arXiv-ID • DOI • Scopus-ID • Pubmed-ID • ISSN
- 8. 8 Distribution of arXiv Articles Over Time: q-bio Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 9. 9 Distribution of arXiv Articles Over Time: astro-ph Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 10. 10 Results of the Data Collection Process Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics 36.1% 50.0% 81.5% 29.8% 76.8% 86.1% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% docs with DOI on arXiv docs with DOI docs found on Mendeley q-bio (1992-2014; n=14,195) astro-ph (2009-2014; n=81,814) Change in the metadata lookup
- 11. 11 Findability on Mendeley: q-bio (1992-2014) 11Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics Total documents 14,195 Documents with DOI & arXiv-ID 7,100 (50.02%) Documents with arXiv-ID 7,095 (49.98%) Documents retrieved on Mendeley with DOI 6,492 (91.44%) Documents retrieved on Mendeley with arXiv-ID 4,896 (69.01%) Documents retrieved on Mendeley with arXiv-ID 5,414 (76.25%)
- 12. 12 Findability on Mendeley: astro-ph (2009-2014) 12Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics Total documents 81,814 Documents with DOI & arXiv-ID 62,800 (76.8%) Documents with arXiv-ID 19,014 (23.2%) Documents retrieved on Mendeley with DOI 54,093 (86.1%) Documents retrieved on Mendeley with arXiv-ID 12,812 (67.4%) Documents retrieved on Mendeley with arXiv-ID 45,586 (72.6%)
- 13. 13 Findability on Mendeley & DOI Coverage: q-bio Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 14. 14 Findability on Mendeley & DOI Coverage: astro-ph Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2009 2010 2011 2012 2013 2014 findability DOI availability
- 15. 15 Identifier Frequency and Mean Readership on Mendeley Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics arxiv doi scopus pmid issn q-bio (n=11,570) frequency 10,351 (89.5%) 8,321 (71.9%) 8,409 (72.7%) 5,477 (47.3%) 8,119 (70.2%) mean readership 20.4 25.4 25.4 32.4 25.9 median readership 9 13 13 19 14 astro-ph (n=70,459) frequency 65,719 (93.3%) 61,667 (87.5%) 58,780 (83.4%) 1,454 (2.1%) 60,985 (86.6%) mean readership 7.2 7.3 7.5 20.6 7.3 median readership 5 5 13 5 6 StdDev: 33.4
- 16. 16 Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics Distribution of Disciplines on Mendeley (March 2011) [Kraker et al. 2012]
- 17. 17 Identifier Combination Frequency on Mendeley Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics 39.2% 22.8% 18.2% 4.4% 4.1% 2.7% 0.0% 0.0% 0.0% 8.5% 1.3% 10.2% 74.5% 0.0% 0.0% 4.5% 2.6% 1.7% 1.5% 3.7% 0% 20% 40% 60% 80% 100% arxiv-doi-issn-pmid-scopus arxiv arxiv-doi-issn-scopus doi-issn-pmid-scopus arxiv-scopus doi-issn-scopus doi-arxiv-issn doi-issn doi-arxiv Other astro-ph (2009-2014; n=81,814) q-bio (1992-2014; n=14,195)
- 18. 18 Conclusions • As expected, crowd-sourced systems show big differences in identifier coverage and distribution crowd-sourced data needs to be amended automatically • When retrieving arXiv articles from Mendeley, we were able to obtain more articles using the DOI than the arXiv-ID • BUT: a single arXiv-ID is the second most popular identifier combination on Mendeley. This suggests that pre-prints are being read – if at a lower level • There is, however, a certain time lag concerning the adoption of articles from arXiv on Mendeley Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 19. 19 Conclusions • The distribution of identifiers in a collection of papers gives hints at the nature of papers in this collection • As with citations, field normalization will be essential for cross-comparison of altmetrics scores Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 20. 20 Limitations & Future Work • Choice of arXiv as primary data source introduces a bias • We only looked at two disciplines from natural sciences • Using a random sample of 381 articles from Web of Science, Zahedi et al. (2014) report that they were able to retrieve only 47.7% of articles on Mendeley using the DOI or the title. Expand this study to further disciplines and fields using more data sources Look more deeply into the reasons for disciplinary and tool-related differences Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics
- 21. 21 Open Source Crawling Framework Know-Center GmbH • Research Center for Data-Driven Business and Big Data Analytics https://github.com/Bubblbu/crawling-framework
- 22. 22 b SAVE THE DATE: i-KNOW Conference Special Track on Science 2.0 & Open Science 21-23 October 2015, Graz, Austria Submission Deadline (extended): 22 June 2015 (Abstracts due: June 8)