FA678945/github-issues-flask-augmented

🗂️ Dataset Card: GitHub Issues Dataset for Flask (Augmented)

Author: FA678945 Version: 1.0 (2025) License: CC BY 4.0 Source: pallets/flask GitHub Repository

📘 Dataset Summary

This dataset contains a curated collection of public GitHub Issues from the Flask web framework repository (pallets/flask). It has been augmented with derived metadata to support educational and research applications including sentiment analysis, text classification, and semantic search.

Created for BSBDAT501, ICTDAT401, and BSBCRT404 assessment tasks, this dataset demonstrates ethical, compliant, and reproducible dataset creation using a public API.

📚 Supported Tasks & Benchmarks

This dataset is suited for:

Sentiment Analysis

Text Classification (bug vs non-bug, topic categorisation)

Semantic Search using FAISS

Keyword Extraction / Text Cleaning

EDA & NLP Demonstrations

Curriculum-Based Training for Dataset Creation and Documentation

This dataset is not intended for real-world decision-making or automated actions affecting individuals.

📡 Data Source GitHub API Endpoint https://api.github.com/repos/pallets/flask/issues

Collection Method

Data was retrieved using Python’s requests library:

import requests

url = "https://api.github.com/repos/pallets/flask/issues" response = requests.get(url) data = response.json()

Only public issues were collected. No authentication, private repos, or sensitive data were involved.

🛠️ Data Fields Raw Fields Field Type Description id int GitHub issue ID title string Issue title body string Issue body text state string “open” or “closed” labels list List of label objects user string GitHub username (public) createdat string ISO timestamp updatedat string ISO timestamp Augmented Fields Field Type Description titlelength int Character count of title issuelength int Character count of body numlabels int Number of labels applied isbug bool True if issue text contains “bug” createdmonth int Month extracted from timestamp textclean string Lowercased, punctuation-removed text sentiment float/label Optional sentiment score/class keyword_present bool Checks for keywords like “error” or “exception” 📁 Dataset Structure

Each record represents one GitHub Issue. The dataset is provided as a structured table (e.g., JSON, CSV, Parquet, HF Dataset format).

🧪 Dataset Creation Process

1. 1.Data Extraction

Retrieved via GitHub REST API

Limited to public issues

Includes only publicly visible usernames

1. 1.Cleaning

Removed empty descriptions

Normalised whitespace

Lowercased and stripped punctuation

1. 1.Augmentation

Added fields for length, sentiment, and keyword features to support NLP tasks.

1. 1.Validation

Checked for malformed timestamps

Ensured all fields had consistent types

Verified no personal/sensitive data included

🔐 Ethical & Legal Considerations Compliance with Policies

This dataset complies with:

Koorliny Kaatijin AI Use Policy

Only public data used

Full transparency and documentation

Data Access & Classification Policy

Classified as Public

Privacy Policy

No personal or sensitive information collected

Information Security & Acceptable Use Policy

Stored on an approved cloud repository (Hugging Face)

Data Retention & Disposal Policy

Dataset may be removed after academic use

Privacy

This dataset uses only publicly available GitHub usernames, which are explicitly non-sensitive.

No email addresses, private repositories, or personal data were collected.

Copyright & Licensing

Original GitHub content is subject to GitHub Terms of Service.

Augmented dataset is released under CC BY 4.0.

Attribution is required (see below).

⚠️ Limitations

Informal writing style in issues

Small sample size

Heuristic fields like is_bug may be noisy

Label and topic imbalance

Possible contributor-style bias

Not suitable for production ML systems.

🌐 Intended Uses

This dataset is intended for:

Classroom/academic demonstrations

Text preprocessing and NLP tutorials

Training on dataset creation & documentation processes

Experimenting with FAISS search

Exploratory data analysis

Not suited for:

Automated decision-making

High-risk AI applications

Profiling individuals

📜 Licensing Dataset License

Creative Commons Attribution 4.0 (CC BY 4.0)

Attribution Requirements

Please credit:

Original Source Data: pallets/flask GitHub repository

Issue Authors & Contributors

Dataset Creator: FA678945 (2025)

Dataset URL: Hugging Face link

🧾 Citation FA678945 (2025). GitHub Issues Dataset for Flask (Augmented). Hugging Face Datasets. Source data from the pallets/flask GitHub repository. Licensed under CC BY 4.0.

📈 Future Work

Possible enhancements:

Embeddings for FAISS (Task 3)

Larger set of issues

Label-based classification tasks

Triage prediction model

Topic modelling