Impact of the COVID-19 pandemic on academic productivity

<p>This data set was used to analyse changes in&nbsp;academic productivity due to the COVID-19 pandemic.</p> <p>Here we describe how those data were collected. We downloaded the <a href="https://www.kaggle.com/datasets/Cornell-University/arxiv">arXiv dataset</a>&nbsp;hosted by Kaggle on 2022 June 6. This dataset is a JSON file (named `arxiv-metadata-oai-snapshot.json`) compressed into a zip file. We extracted this file and then executed&nbsp;`process.py` to create the following three files:</p> <p>- `metadata.json`: titles, abstract, and affiliations of authors</p> <p>- `authors.json`: author names</p> <p>- `records.csv`: arXiv identifier, categories, number of authors and affiliations, abstract length</p> <p>&nbsp;</p> <p>We excluded pre-prints from 2022 June when processing files. The `process.py` script also adds a &#39;dummy&#39; line to the `records.csv` to make Python load arXiv identifiers as strings. These three files (`metadata.json`, `authors.json`, and `records.csv`) were used for this analysis.</p>