1 / 11

VIVO Expert Finder Update

M. Conlon June 10 , 2014. VIVO Expert Finder Update. Why an Expert Finder?. Collaboration is the future of scholarship Finding experts can be difficult Cost/benefit for development is changing with new tools We have the data. How: Architecture. The Data.

ania
Télécharger la présentation

VIVO Expert Finder Update

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. M. Conlon June 10, 2014 VIVOExpert Finder Update

  2. Why an Expert Finder? • Collaboration is the future of scholarship • Finding experts can be difficult • Cost/benefit for development is changing with new tools • We have the data

  3. How: Architecture

  4. The Data • Papers -> Concepts, Papers -> People (Experts) • Concepts on papers – PubMed and other indexing services associate concepts with papers • People on papers as authors • Associate people to concepts on which they have published • At UF: • 25,878 people with papers • 18,520 concepts • 20,408 papers with concepts

  5. A Detour Through Histograms • We’d like to see the data. We start with simple histograms • Good for learning SPARQL, JSON, Ajax, D3 • Good for learning about our data • Demo Ajax histogram for org types • Demo JSON Histogram for org types • Concept counts and its SPARQL query • Concepts per paper and its SPARQL query

  6. Connect Concepts to People via Papers Tumor Suppressor Protein p53 William Stratford May, Jr “Strat” May has written two papers on Tumor Suppressor Protein, p53

  7. So we will need a lot of data • All concepts, all people, all papers, connected = VIVO! • But we need to have some pre-counted JSON data ready for web requests = a concordance • So we use Python to process a SPARQL query concordance result set and create a JSON concordance Oops. Too big. We need another approach So we use Python for each concept to find concurrent concepts and authors Oops. Too big. We need another approach So we use Python to subset the concordance and select “dense” connections Now we’re ready

  8. What – bi-modal force directed graph of concepts and people

  9. The real thing • The concordance subset process has trimmed the concordance globally to only concepts that appear two or more times, co-occurrence of concepts must be at least two. Authors must have published at least two papers regarding the concept. • For a specific concept, find all the co-occurring concepts, and for each concept, find all the people who have published on that concept. Produce a JSON file ready for display. See make_graph_json.py • The resulting JSON is then shown using a force-directed bi-modal graph. See Demo

  10. Next Steps • Size concept nodes on number of papers • Size author nodes in two rings – inner ring size based on number of papers on the concept. Outer ring on the number of papers total • Weight connects on the number of papers in the connection • Provide hover navigation for concepts to the concept in VIVO, people to their profile in VIVO, links to a display of the papers comprising the connection (with links for each paper to VIVO) • Provide user controls for trimming and positioning the network • Provide a means for accessing all the concordance data without having to generate a new JSON file via Python • Provide recentering – double click on concept or author to redraw the network around the select concept or author • Provide output for identifying people of interest (experts)

  11. Repos • Code • http://github.com/mconlon17/mconlon17.github.io • Some VIVO Things web site • http://mconlon17.github.io • Demo • http://mconlon17.github.io/angular/bimodal.html

More Related