---
license: cc0-1.0
---
# 🧠📚 ArXiv Research Papers Metadata – Finance (q-fin) & Computer Science (cs)

This repository provides two extensive and high-quality metadata datasets from [arXiv.org](https://arxiv.org), covering research papers in:

- 💸 **Quantitative Finance (q-fin)**  
- 💻 **Computer Science (cs)** (all subcategories)

Each dataset includes structured metadata fields and **direct PDF download links** for every paper. Together, these datasets serve as valuable resources for machine learning research, NLP model training, academic search, and more.

---

## 📦 Datasets Included

### 1. 🏦 Quantitative Finance (`q-fin`)
- 📈 ~Thousands of finance-related papers
- ✅ Focused on high-quality papers in areas like:
  - Financial modeling
  - Market analysis
  - Algorithmic trading
  - Risk management
- 🎯 Great for domain-specific LLMs, finance QA, or economic text mining

### 2. 💾 Computer Science (`cs`)
- 📊 Includes **500,000+** papers across **all CS subfields**
- Covers:
  - Artificial Intelligence (cs.AI)
  - Machine Learning (cs.LG)
  - Natural Language Processing (cs.CL)
  - Computer Vision (cs.CV)
  - Software Engineering, Theory, Networks, and more
- ⚙️ Ideal for LLM pretraining, summarization, or scientific search tools

---

## 📑 Features

- Direct PDF links to full papers (`pdf_link`)
- Rich metadata:
  - `title`, `authors`, `abstract`
  - `categories`, `submission date`, `arXiv ID`
  - Optional: `is_scraped` flag for tracking downloads

---

## 🧠 Use Cases

- Pretrain or fine-tune LLMs on domain-specific scientific texts
- Build citation tools or academic paper search engines
- Create offline datasets of research papers
- Academic summarization, QA, paper classification, and recommendation systems

---

## 📜 License

This dataset is released under the **[CC0 1.0 Public Domain Dedication](https://creativecommons.org/publicdomain/zero/1.0/)**.  
> You are free to copy, modify, distribute, and use this dataset for any purpose — even commercially — without needing permission.

---

## 📂 Dataset Format

Each row in the CSV files corresponds to a paper.

| Column         | Description                                       |
|----------------|---------------------------------------------------|
| `id`           | arXiv ID (unique identifier)                      |
| `title`        | Paper title                                       |
| `authors`      | Author names                                      |
| `abstract`     | Abstract text                                     |
| `categories`   | arXiv subject classification (e.g., `cs.CL`)      |
| `pdf_link`     | Direct link to the paper’s PDF                    |
| `is_scraped`   | (Optional) Indicates if the PDF was downloaded    |

---

## 📁 Files Included

- `qfin_metadata.csv`: Metadata for Quantitative Finance papers
- `cs_metadata.csv`: Metadata for Computer Science papers
- `qfin_downloader.ipynb`: Script to batch download finance paper PDFs
- `cs_downloader.ipynb`: Script to batch download CS paper PDFs

---

## 🙋‍♂️ Maintainer

Created and maintained by **[G Gowtham]**.  
code for gether metadata and pdf's form arXiv avlible at [git hub](https://github.com/Gowthamcopilot/arxiv_papers) 
Feel free to contribute, ask questions, or report issues!

---

## 🙌 Acknowledgements

Thanks to [arXiv.org](https://arxiv.org) for making academic research openly accessible.