I am working on some project, in which I would like to find articles where they use some computer code to treat data, and the used datasets and code is totally available (github, personal website, etc). So I go to Scopus, Google Scholar, ISI Web of Science, etc, but I do not find a good strategy for this purpose. I try to add keywords to the search like "code available" or "github", but it does not always yield accurate results. I also know that some journals force papers to give datasets and code, but do not know how to find these exact journals in my domain. So, what is the best strategy for finding articles that share code and datasets?
-
4paperswithcode.com. It contains the publications and the corresponding code. Most top computer science conferences ask for relevant supplementary material, which includes code. Many papers (atleast in Computer Vision/Machine Learning) which are published in top venues contain links to the code as well.– JihadiApr 28, 2020 at 7:27
-
Many professors also maintain public git repos, which contains the code for their publications. You can search on Github or go to the homepage of the professor to find the link.– JihadiApr 28, 2020 at 7:32
-
1You can also email the author directly. In my experience, many are willing to share it.– user760900May 4, 2020 at 0:18
-
What field are you in?– Richard EricksonDec 4, 2020 at 14:25
-
try github search– Mike LiuJul 26 at 2:02
3 Answers
paperswithcode.com. It contains publications and the corresponding code.
Most top computer science conferences ask for relevant supplementary material, which includes code. Many papers (atleast in Computer Vision/Machine Learning) which are published in top venues contain links to the code as well.
Many professors also maintain public git repos, which contains the code for their publications. You can search on Github or go to the homepage of the professor to find the link.
zenodo.org is a repository that enables one to index code hosted on Github. Most probably you would find a Zenodo DOI associated in the supplementary material. If not, I've seen some journals like RSC Analyst, putting the link along with the research article on their site.
Journal of Statistical Software and R Journal both require replication materials for all papers.