Collaborations Workshop 2019 (CW19) #CollabW19          2019-04-01 to 2019-04-03

Group 1 - CI1-CW2CC

Reporter

Rachael Ainsworth - rainswor@gmail.com

Participants

Rachael Ainsworth (Chair), Simon Hettrick (scribe), Claire Wyatt, Gabriel Hanganu, and Chris Mentzel

This document should be used to capture the information for a Collaborative Session / Hack Day Idea. (The total amount of text should ideally be between 100-300 words and you can include a diagram or two). The document should be no larger than two pages of A4. Don’t delete the details at the top of the document but you should delete all the hint text once you no longer need it.

Agenda & info: https://software.ac.uk/cw19/collaborative-ideas-and-hackday-ideas

Context / Research Domain

Please describe the context or research domain to which the problem applies

The success of the RSE community grew from a small group of people identifying a new role in academia that was important, growing and almost completely unrecognised. We want to identify new roles so that we can help them repeat the same success.

Problem

Description of the problem you are trying to solve

There are a number of roles in research that make a significant contribution to research, but go unrecognised. Once these roles have been identified, we can name them and help them gain recognition - the difficulty lies in identifying the roles.

Solution

Explanation of the solution to the problem you have identified

We’ll use job sites (probably jobs.ac.uk) to identify new roles:

  1. Scrape currently available job adverts
  2. Separate the data into different categories based on job title
  3. Start with postdoctoral positions (will require some simplification of the different titles used) and conduct a word frequency analysis on the associated job advert text to gain insight into the described roles
  4. Attempt to identify the “key responsibilities” then analyse this data (potentially using NLP methods) to simplify responsibilities which can then be compared across job adverts
  5. Identify the most similar roles (say, top five across each role) then review them manually to identify any names, terms or language that identifies the role (or is being used to identify it)
  6. Take the identified names, terms or language and search through all job adverts to identify any roles that could be associated with it.
  7. If there’s a sufficient number of similar of roles identified, use the information to build a profile of the role. And name it, if necessary.
  8. Publish all of the roles we discover and attempt to get people to rally around them.

Additional analysis

  1. Take the roles we’ve identified
  2. Review the balance of fixed term and permanent contracts
  3. If the role is mainly associated with permanent contracts, this could indicate that the role is accepted, if they are mainly fixed term, then it could indicate that the role is unrecognised
  4. Conduct a frequency analysis on job titles
  5. Identify job titles, other than postdoc, that are numerous and do not appear to be recognised
  6. Conduct the main analysis, as described above on postdoctoral jobs, but with these new job families.

Diagrams / Illustrations

You can include one or two diagrams in this section. Please ensure you have the right to use the image(s), and include an attribution if applicable.