GitHub OSDB Social-Technical Network Reference Relationship Dataset 2023

The GitHub OSDB Social-Technical Network Reference Relationship Dataset 2023 is a specialized resource for studying the technical dependencies, collaboration patterns, and ecosystem evolution within the Open Source Database (OSDB) domain. This dataset is built upon the theoretical foundation of Reference Coupling (Blincoe et al., 2019) and the improved IREL method (Liu et al., 2022), focusing on extracting and analyzing cross-project reference relationships from GitHub event logs. # Key Features 1. **Domain Focus**: Exclusively targets open source database management systems (DBMS) projects, including relational databases (e.g., MySQL, PostgreSQL), NoSQL databases (e.g., MongoDB), time-series databases (e.g., ClickHouse, InfluxDB), and other database technologies. 2. **Data Sources**: Derived from OpenDigger's ClickHouse database, which is built on GHArchive and provides comprehensive GitHub event log data. The dataset includes events from 2023, with a focus on Issues, Pull Requests, and Commits. 3. **Data Collection Methodology**: Constructed using the GH_CoRE (GitHub Collaboration Relation Extraction) tool, which employs: - Multi-pattern matching strategies for reference identification - Strict entity existence verification to ensure data quality - Comprehensive error handling and logging mechanisms - Caching optimization for efficient data processing 4. **Data Structure**: Based on a heterogeneous information network (HIN) model, the dataset organizes information into: - Node types: Actor, Branch, Commit, CommitComment, Gollum, Issue, IssueComment, PullRequest, PullRequestReview, PullRequestReviewComment, Push, Release, Repo, Tag. - Edge types: Reference (core innovative feature), EventAction, Contribution - EventAction records containing entity relationship information, event metadata, and contextual information 5. **Quality Assurance**: Implements rigorous data cleaning processes including: - Deduplication strategies (event deduplication by Event ID, entity deduplication by global ID generation) - Reference entity resolution and validation (direct parsing, pattern matching, API verification) - Data validation mechanisms (Token Pool mechanism, pattern improvement, quality standard checks) - Context preservation and link decay handling # Use Cases This dataset is suitable for: - Open source software ecosystem research - Socio-technical network analysis - Technical dependency mining - Database domain evolution studies - Collaborative pattern analysis in software development # Access and License The dataset is available for download via Google Drive and campus intranet access. All data is sourced from GitHub's public API and only includes events from public repositories, complying with GitHub Terms of Service. The dataset is released under the Creative Commons Attribution 4.0 International. # Associated Tools - **GH_CoRE**: A PyPI-published tool for extracting reference relations from GitHub logs, which can be used to extend or reproduce the dataset. See [gh-core](https://pypi.org/project/gh-core).