Project name: Exploring Disciplinary Differences in Software Mentions

Project description

Project slides

This project was part of the Chan Zuckerberg Initiative on "Mapping the Impact of Research Software in Science". In this project, we are interested in studying the following questions:

• What is the distribution of publications mentioning (or not) software across disciplines?
• How is different software used by researchers across their publications?
• What is the ‘proximity’ of scientific publications to the use of software? (ongoing)

Methodology

We conduct scientometric analysis of publications mentioning software to match software mentions with papers, authors, and disciplines.

Datasets

• CZI mentions dataset (1.7 million PubMed IDs)
• [OpenAlex] (https://openalex.org/)
• [SoftwareKG] (https://zenodo.org/records/3715147)

Software/Tools

• Google BigQuery (InSySPo project - Brazil)
• Databricks
• VOSviewer
• R
• Python

Data collection

Match CZI software mentions and SoftwareKG mentions with OpenAlex publications (DOI, PMCID)

Software name disambiguation in CZI dataset

There were software names in the CZI dataset that were not disambiguated. We used fuzzy matching to identify the "similar" software names to merge them before plotting our networks.

Findings

Top softwares per discipline

Software mentions per discipline across time

Software mention networks

Using the CZI dataset (1.7 million publications)

Using the KG dataset

Software network differences across contrasting disciplines

Future work

Software dependency per domain

Software dependency domain comparison

Contributers

• Alexy Khrabrov
• Frank Krüger
• Fuqi Xu
• Huimin Xu
• Puyu Yang
• Rodrigo Costas
• Shahan Ali Memon