This artifact contains the dataset used in the paper Understanding the Role of External Pull Requests in the NPM Ecosystem.

Script

We provide 3 script files

1. preliminary_A1_2.ipynb for the preliminary A1 and A2
2. preliminary_A3.ipynb for the preliminary A3
3. rq1.ipynb for RQ1

Note that RQ2, we used seperated spreed among three authors and count the result.

Dataset

The appendix contains our quantitative and qualitative datasets, shown in Figure 1 in the paper. The dataset are provided via this repository and zenodo for the large dataset* (file size more than 100MB).

There are seven data files as described below:

1. filtered_dataset_with_label.csv - A filtered dataset contains 945,291 PR.*
2. external_PR_dataset.csv - List of the External PR used in preliminary study, RQ1, and RQ2.*
3. internal_PR_dataset.csv - List of the Internal PR used in preliminary study, RQ1, and RQ2.*
4. bot_PR_dataset.csv - List of the Bot PR used in preliminary study, RQ1, and RQ2.*
5. attention_label_ranking.json - The attention labels of each PR type with count using in RQ1.
6. ranking_label_bot.csv - The labels appearing in the Bot PR are already ranked and be used in RQ1.
7. ranking_label_external.csv - The labels appearing in the External PR are already ranked and be used in RQ1.
8. ranking_label_internal.csv - The labels appearing in the Internal PR are already ranked and be used in RQ1.
9. sampled_dataset_external_PR.csv - Result of manual classification of the External PRs used in RQ2.
10. sampled_dataset_inside_PR.csv - Result of manual classification of the Internal PRs used in RQ2.
11. sampled_dataset_bot_PR.csv - Result of manual classification of the Bot PRs used in RQ2.

Note that, 1-4 are the large dataset files.