Week 4

Minutes:

• Meeting started with Abhinav introducing the agenda.
• Justin asked Dr. Rothermel about i10-index, h-index, and self-citations. Self-citations are not a problem to Dr. Rothermel (it can sometimes show a work’s relevance), but i10-index and h-index will create useful queries.
• Abhinav clarified Dr. Rothermel’s query case by the year.
• Tyler and Gage went through our API Endpoints, asking about if all the routes cover the query cases.
• Dr. Rothermel brought up adding a faculty’s position or field to our database for queries.
• Possible tags include: rank, area, a temporary “candidates” tag etc.
• The set could be filtered out in the front-end by whatever query type through the same API endpoints.
• He wants the option to tag based on a list of authors whose position data is null.
• Filtering by multiple tags is desired, too.
• Gage showed Dr. Rothermel a test coverage report, where our current code has 93% line coverage.
• Carter asks to go over his desired visualizations. There are three basic outputs.
• Box-plot is obvious, showing professor to citation in a mostly visual way.
  • Carter brings up the case of having a box for each professor over years.
  • Box plot could also show in year x, a number of cites of a professors papers over time, i.e. showing the impact of a professors work on that year.
    • This information would not immediately be available, and would needto be gathered over time.
  • Another option would be for a professor in a year, the number of citations of papers published in each year, i.e. the more impact of a paper in a given year.
  • Gage shows that google scholar uses the citations made of a faculty member’s work in a year, but the meaning is particularly vague.
  • Outlier removal, in extreme cases, should be considered.
  • Line or bar graph with time (years) to faculty member. This has a couple interpretations, so more research is needed. Line plot for articles’ citations over time.
  • Carter asks about the usefulness of non-cited articles.
  • Spreadsheet output is the most basic return. There seem to be 2 useful variants so far:
    • A column corresponds to a professor, and for that the citations per article per person should be an option. This may include a paper title column.
    • A raw data csv file would be useful for managing the data in excel or reporting to interested parties.

Action Item	Person Responsible	Due Date
Identify what the Google Scholar graph considers	Gage/Tyler/Abhinav	Monday (2/14)
Update on scraping	Carter/Justin	Monday (2/14)