1 of 10

eLife Sprint 2020 Report

@codeisscience

github.com/codeisscience

2 of 10

Goal

“Whilst it is generally agreed that scientific research needs to be peer-reviewed as part of the publication, this stipulation doesn’t always extend to peer-reviewing code. We aim to build infrastructure that clearly lists not only journal policies with regards to code artefacts, but also compliance with journal policy.”

3 of 10

Journal leaderboard

Initially sounds easy - show which journals have open policies and which journals comply with the policies.

6 of 10

Implementation

Frontend

To be implemented in something common and familiar with a lot of developers such as React JS.
A modern design for the UI /UX

Backend

We’ve decided to go with something common with a lot of developers such Python
The DB will be based on open source tools.

7 of 10

A simple DataBase scheme

8 of 10

How do we rank Journals? Discussion

How do you identify papers with computational methods but no code?

Our best guess for now - keywords.
“Custom scripts” seems especially likely to have computational methods but no code sharing

Should there be standardisation of code-sharing format for journals (time for a working group?) - as even when code is shared...

Sometimes code is in “data availability”
Sometimes it’s “methods”
Sometimes it’s free text
Sometimes it’s a PDF in the supplementary materials 😭

9 of 10

Ways you can help out

Interested in taking this further?

Visit our GitHub and look at issues - maybe one is for you?

Add more keywords in our Sprint Google Doc - especially field-specific keywords

Sign up to our mailing list - we’ll be running community calls to continue our work! (Visit http://www.codeisscience.com/ - sign-up link in footer).