User Input for Search Engine Workflow

[ChristinaB]

Steps to create a graph network connected by a search term instead of an author?
Current connections create edges from a list of people.

New type of graph network: create edges from a list of terms: tools or topics
Humans, leaders, and organizers determine the short list of search terms and publish visualizations in report.

• Promote and recruit collection of updated bibtex files (manual)

• Request permission to scrape OrcID and Github and data repository and get a list of IDs and repositories

• Add code to scrape OrcID

• String manipulation to reconcile names to develop consistent ID connection for nodes

• Scrape HydroShare (to do other data repositories)

• Scrape Github based on input 'Organization Name' requires permission and Admin access or key generated to access the API

• Organizer/Coordinator User Inputs = bibtex file, list provided by network members

• Natural language processing on abstracts and readme and all metadata fields. (See github.com/hydroshare Discover search engine does this, for example). Output = List of terms

• Mapping of keywords corresponding to each term/keyword.

• View graph network with color coded nodes and links by tags in the search. MVP = pick 5 keywords, make a graph network view that provides a filter with check box that turns on and off like Github, HydroShare, Journals

• Print selection of filter to make a table and report by tags