2021-03-25

Background for ABM project

Goal:

  1. Replace the previous ABM i.e. Annual Bibliometric Monitoring / Årlig Bibliometrisk Uppföljning at KTH
  2. Further develop the ABM based on input from this group

Summary:

Our previous demo meeting was held 2020-11-10. After christmas vacations we have worked on Sprints 24 to 31 (current).

Netiquette:

Please use Zoom, raise hand for questions or write in the chat, mute if not speaking, and think about questions for the Q&A session at the end.

Agenda

  • Recap from previous meetings - What feedback has been given so far in the project?

  • Progress - What is new since last meeting?

  • A demo presenting new work (based on 2020 data update)

Short break, with possibility to look at applications

  • Some interpretations and comparisons between data sources and models for ABM

  • Update about other efforts (data integration, APIs etc)

  • Your questions and feedback

Demo group feedback

You have provided many suggestions about steps forward, beyond the “old” ABM application. Here is a rough attempt to summarize input since the project start into some themes:

  • One theme has been Scopus data and its inclusion in ABM

  • Another theme has been about presenting results at division level in an author-based variant of ABM that focuses more a “current view”

  • User Experience improvements with closer alignment with KTH Style on the web

  • Addition of Open Access data/reports including Altmetric integration

Feedback Recap (continued …)

  • Data integration against other (non-WoS) sources
    • Requests regarding Microsoft Academic, Semantic Scholar, Google Scholar etc and ORCID matching (added in Kontarion and DiVA package)
    • Coverage for conference proceedings in CS (partially addressed by Scopus)
    • Custom “divisions” with arbitrary researchers (a POC experiment made in separate app)
  • Another theme has been ideas about new ways to interactively access, view and explore the ÅBU data
    • Data Quality and Curation aspects, for example for usage of HR / researcher data from KTH
    • Raw data access, APIs (mostly non-public APIs due to usage rights for WoS and Scopus data)

Progress - What is new during Q1?

  • Data update with updated data for 2013-2019 including RAE data delivery

  • Data integration for Scopus data added

  • Addition of author-based report with output at divisional level based on KTH HR-data

  • Brief interpretation and comparison of results based on new data for 2020

  • New Open Access report included - what does it say?

  • Other related efforts (non-app-specific)

Demo

Interpretation of results

  • How does results based on Scopus and WoS differ?
  • Comparison organizational-based ABM vs author-based
  • What patterns can we see? Trends over time?
  • Highlights and deviations

WoS vs Scopus

  • Additional help in interpretation, e.g. corroborating trends
  • Differences in coverage, depends on subject field and publication profile
  • Field-normalization done against different standardization groups
    • what does different indicator values mean?
    • alternative ways to field-normalization
    • can interact with subject profile, especially for narrower selections (e.g. divisions)
  • Other benchmark needed(?)

WoS vs Scopus, citations

  • WoS is 3 year citation window, Scopus is open citation window
  • Conference papers included in both
  • Neither include self-citations

WoS vs Scopus, normalized citations

  • Conference papers included from Scopus (FWCI), not from WoS (Cf)
  • Papers are compared to other papers in the same data source
  • Cf is normalized by KTHB on Web of Science subject categories
  • FWCI is normalized by Scopus

WoS vs Scopus, journal impact

  • Conference papers included from Scopus, but many have no SNIP
  • In general, journal indicators from Scopus much higher than those based on WoS
  • KTH average relatively high, based on both data sources

Regular ABM vs author-based version

  • Different scope of applications
    • Regular ABM: selection based on links between units and publications (address affiliations)
    • author-based: selection based on current staff and all their KTH publications
  • Probably biased selection of staff categories, due to levels of turnover
  • In general, lower volumes in staff-based version
  • Citation indicators with varying effects between selections

Regular ABM vs staff-based, publications

Regular ABM vs staff-based, Cf

Going forward

Going forward

  • Focusing on additions to collaboration data
  • Suggestion:
    • Add coarse overview of collaborating organizations (as table) to ABM
    • Further collaboration data presented through separate application
  • Working more in practice with new HR-data resource

Other efforts

  • Efforts made are available at KTH-Library@GitHub, including bibliomatrix now with an API and Static site generation tool used for deploying author-based ABM

  • New python tools bibutils for Data Quality and Curation

  • A new R package diva with Data Quality/Curation tooling for the KTH corpus (publications and authors)

  • A new R package cordis provides data access to EU data for research projects, including a new data-only repository ’cordis-data`, piggybacking on GitHub, using it for storing releases of versioned data publicly.

  • Integration with HR data (extract to SFTP site) flowing into modern object storage (using https://min.io/).

Questions and feedback

Questions and Answers

Please provide your input!

  • Do you think the what was presented in the demonstration meets your needs for Scopus data in ABM?
  • What advantages and disadvantages do you see with introducing these additional dimensions into ABM (more data sources, author-based variant)?
  • What needs do you see for presenting collaboration and co-publication data in ABM? Summary tables, networks, author-networks etc.
  • Other questions from the Zoom chat
  • Suggestions and comments

Thank you for attending!