Plum analytics - acs -- 2013-09-09
Download as PPTX, PDF3 likes1,635 views
This document describes Plum Analytics, an altmetrics tool that tracks the impact of scholarly works across a wide range of online sources and metrics. It summarizes some limitations of traditional citation analysis and outlines Plum's capabilities, including tracking over 20 types of scholarly artifacts, aggregating metrics from over 20 sources, building profiles of authors and groups, and visualizing impact through tables and charts. The goal is to provide a more comprehensive view of scholarly impact and influence beyond just citations.
Plum analytics - acs -- 2013-09-09
- 1. ? Andrea Michalek, Co-founder andrea@plumanalytics.com @amichalek @plumanalytics Plum Analytics: An altmetrics tool for determining impact in the chemical sciences 1
- 2. Management Team Andrea Michalek – Co-Founder and CEO • Helped launch dozens of early stage companies • Director of Technology for Summon® -- $15M scholarly search used in 450 libraries worldwide Mike Buschman – Co-Founder and Chief Product Officer • Program Manager for Microsoft Academic Search and Book Search—both launched in 2006 • Director of Product Management for Summon® 2
- 3. Current State of Scholarly Measure 3
- 4. Well-known Problems with Citation Analysis • Self-citations • Citations added because of pressure by editors • Negative citations • Only about 30% of influences are cited* • “Informal” influence isn’t measured through citations • Secondary sources • Review articles “take” citations away from reviewed articles • Matthew effect – the rich get richer * MacRoberts and MacRoberts, 2010 4
- 5. 5 Citations are only a small fraction of how a paper is reused 5 0.3% 100% 22.3% Article-Level Metrics collected October 10, 2012 for 62,536 PLOS Papers
- 6. The world has changed 6
- 8. Different categories of impact •Citations •Usage •Captures •Mentions •Social Media 8 Altmetrics
- 9. Tracks 20+ Types of Artifacts • Articles • Blog posts • Book chapters • Books • Cases • Clinical Trials • Conference Papers • Data Sets • Figures • Grants • Interviews 9 • Letters • Media • Patents • Posters • Presentations • Reports • Source Code • Theses / Dissertations • Videos • Web Pages
- 10. Associate Artifacts 10 Artifacts can be associated with a researcher and/or a group Each section can be expanded to fill in details.
- 11. Largest List of Metrics Sources • Amazon • Bit.ly • CrossRef • Delicious • Dryad • dSpace • ePrints • Facebook • Figshare 11 • Github • Google+ • Microsoft Academic Search • PLOS • PubMed • Reddit • Research Blogging • Scopus • SlideShare • SourceForge • Stack Overflow • Topsy (Twitter) • USPTO • Vimeo • Wikipedia • Worldcat • YouTube
- 12. PlumX Platform: Built for Big Data • Built for Scale – goal tracking all the world’s researchers, all of their outputs, and metrics for each. • Technology that solves inherent problems around: • Author disambiguation • Identity Resolution • Calculating combined metrics • Harvesting metrics at scale • Experience architecting scalable platforms • E.g., Summon – provides sub-second response time to searching over 1B documents (including full text of both articles and books) 12
- 13. Different categories of impact • Usage • Downloads, views, book holdings, ILL, document delivery • Captures • Favorites, bookmarks, saves, readers, groups, watchers • Mentions • Blog post, news story, Wikipedia article, comments, review • Social Media • Tweets, Google +1s, likes, shares, ratings • Citations • Web of Science, Scopus, Google Scholar, Microsoft Academic Search, patents 13
- 14. 14 Artifact view
- 15. Build User Profiles from Existing Systems • Set of tools that can utilize existing systems. • Examples • Google Scholar: BibTeX • ORCID • VIVO • Institutional repositories: bePress, dSpace, ePrints • Etc. • Our goal: We integrate with what our customers have implemented limit re-entry of data 15
- 17. Tabular View Sortable, table-view of full output. Exportable to spreadsheets or integrated into other applications. 17
- 18. PlumX Tracks Groups • University • Institute • Museum • Department • Lab • Company • Publisher • Journal • Volume • Issue • Collection 18
- 19. 19
- 20. Metrics Preview in Search Results 21
- 21. “Same” Article, Different Places 22
- 22. Metrics for Every Copy 23 • 59 Bitly clicks through PLOS • 21 Bitly clicks through DOI link • 92 Delicious bookmarks to PLOS • 2 Delicious bookmarks to PubMed
- 23. Researcher Who Blogs 24 •Why is this discovery significant? •What are the most important conclusions? •What does our study NOT say?
- 25. Does the # of citations tell the whole story? • Scopus = 2 • Web of Science = 0 • Google Scholar = 8 • PubMed = 1 Photo credit: A. Wayne Vogl and Nicholas D. Pyenson / Smithsonian Institution. 26
- 27. 28 Author / Group Visualization
- 28. Group, Author and Artifact Widgets 29
- 29. Thank You! Try PlumX Today http://try.plu.mx @plumanalytics @amichalek 30 Andrea Michalek andrea@plumanalytics.com 215.280.1805 30
Editor's Notes
- #4: Published articles in prominent journals cite other articles in prominent journals = prestige and tenure.This mechanism of using citations in published journals to determine impact of research was built in the 1960’s.
- #7: As everyone knows, information has moved online.Where printed books and journals were once the only scholarly record, this too has become digital.The current impact measurements for research were defined in the 1960's, when the world was a different place.They are only based at looking at what papers cite other papers in the reference section of a document.
- #20: Categories of metrics
- #22: Compare the metrics showcasing the capturing of immediate impact versus a search on Google Scholar for something like “brain cancer.”
- #23: “Same” article on PLOS, PubMed, and Pitt’s institutional repository.Metrics need to be algorithmicallycombined from ALL places to capture all of its use.
- #24: http://demo.plu.mx/a/eKXLDK8_gEB-iSPG_HLlHVGtDVYDreSTzdY7VST1rmY/Plum’s Identity resolution system, can match different artifacts together by known identifiers (like DOI), aliases of URLs, and proprietary clustering analysis algorithms.This system scales across millions of artifacts, and gives a complete view to the metrics.
- #25: NicholasPyenson blogs about his research.In a recent (2012) Nature paper that was chosen to be one of Discover Magazine’s top science stories of the year.His work has gotten pickup in the popular press and he has blogged about key areas relating to his research.Traditionally, he will not get “credit” for this work.http://nmnh.typepad.com/pyenson_lab/nature-rorquals-organ.html
- #26: Other types of artifacts we track for this research include: BBC article about his work that was liked on Facebook over 700 times and tweeted nearly 500 times A popular blog post about his research An article in the New York TimesThese 2nd level metrics can also be used to help quantify not just the paper he authored, but the reach and engagement of others writing about his work.
- #27: At the time this snapshot of metrics was taken, his work was newly published. It hadn’t been picked up in Web of Science yet at all.However, the metrics surrounding his work give us a leading indicator that it may be impactful research.
- #29: This visualization of a researcher’s output, when weighted by impact, shows which artifacts have been getting the most engagement across their entire life’s work.This is very useful when trying to get an overall sense of what a particular researcher (or lab, or group) is most known for.
- #30: PlumX offers embeddable widgets for Publishers, Institutions and Researchers.We offer both Author level and Artifact level widgets.