aurora network logo

Aurora SDG Bibliometrician Meetup

February 4, 2019

Vrije Universiteit Amsterdam

Although Aurora activities have a clear focus on what the university presidents have defined as the key prioirities, others in the Aurora network also see the potential to exchange information on the issues that are high or their agendas. Meetings in this contexts have been dubbed “birds of a feather” meetings: administrators or academics with similar interests meeting outside the context of the key priorities of Diversity, Societal Impact & Relevance of Research, Student Engagement, and Innovation of Teaching & Learning.

During the Aurora Antwerp Biannual, the team described a working method that balances the workload and the validity of the dataset defining the SDG’s. The main outcomes included slicing the query building in smaller steps; validating a SDG-text-classification model (machine learning) and combining the SDG query publications with high SDG probability publications.

On February 4th from 10:00am – 17:00pm, the bibliometrics team organizes a physical meeting to discuss further procedures in enhancing the SDG analysis: Bibliometrics of  Relevance Instrument. The following items will be discussed during this meeting:

Query Crafting

  • Update: Times Higher Alignment on SDG queries.

Demonstrate progress: the To Do list we have come up with in the Antwerp-meeting:

  • Parse the 240.000 publications (2009-2018) from Aurora universities through the SDG-classifier [Aberdeen, David]
  • Create a list of publications that only all-in one SDG (according to the query method) [Duisburg-Essen, Felix]
  • Match the 25.000 publications (2009-2018) we already found [Duisburg-Essen, Felix]
  • The non-matching publications (N1: where david has found an SDG with >95%, and we have not found), we use to extract additional keywords. [Gothenberg, Lars] (using BibExcel: Ask Susanne; or using “Online Ngram Analyser” or other Term extracting tools)
  • Manually check the publications that match in both models  (N0) (sample of 200) [Antwerp, Raf] (using Rayyan: Ask Lars)
  • The triple positive publications [gold set] are given to David to improve his model [Aberdeen, David]
  • Planning new meeting [VU, Maurice]
  • Next steps

Data Collection

  • SDG Query evaluator & Data collector

Data Visualization

  • Power BI & Elastic Search

 

 

Back to events