📚 Misraj Structured Data Dump (MSDD)

A formatted large-scale Arabic text dataset extracted and filtered from the Common Crawl dumps. This dataset aims to provide high-quality, diverse Arabic textual data suitable for training large language models, especially for Modern Standard Arabic (MSA) and Arabic dialects.

---

📌 Dataset Summary

• Source: Subset from multiple Common Crawl dumps.
• Documents: 23 million documents.
• Timeframe: - 2024 Dump: Dump 10
  • 2025 Dump: Dump 13
• Languages: Primarily Arabic (MSA and dialects).
• Format: Raw text converted to Markdown format.
• Domain Variety: General web content.

---

💡 Usage

📥 Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("Misraj/msdd")

📋 Example Usage

# Access the first example
example = dataset['train'][0]
print(f"UUID: {example['uuid']}")
print(f"Markdown: {example['markdown']}")

---

⚙ Data Processing & Cleaning

The dataset has been carefully filtered and deduplicated to ensure high linguistic quality and to remove noise commonly found in raw Common Crawl text.

✅ Paragraph-Level Filtering

Each paragraph in the corpus undergoes the following checks:

• Character Deduplication: Removal of repeated characters beyond a threshold.
• Word Repetition Ratio: Filtering paragraphs with excessive word repetitions.
• Special Character Ratio: Filtering based on the proportion of non-standard characters.
• Language Identification: Only Arabic paragraphs are retained.
• Perplexity Scoring: Content scored using an in-house KenLM-based model trained on Wikipedia-like pages, Arabic Twitter data, and dialectal text (e.g., Lahjawi), to remove low-quality text.

✅ Document-Level Filtering

Each full document must pass:

• Word Repetition Ratio: Similar to paragraph level, but with different thresholds for full documents.
• Special Character Ratio: Ensures no document is dominated by symbols, code snippets, or garbage text.
• Language Identification: Verifies the document is primarily Arabic.
• Perplexity Score: Documents are filtered based on perplexity thresholds to maintain fluent, natural text.

---

📂 Dataset Structure

The dataset has the following features:

{
  "uuid": {
    "dtype": "string",
    "_type": "Value"
  },
  "markdown": {
    "dtype": "string",
    "_type": "Value"
  }
}

---

🚦 Quality Checks

The dataset quality was validated using:

• In-house KenLM-based Arabic models for perplexity checks.
• Manual inspection of samples.
• A pipeline inspired by OBELICS, with custom enhancements.

---

🔍 Intended Use

This dataset is intended for:

• Training large-scale Arabic language models.
• Research on Arabic NLP, including dialect modeling and low-resource language studies.

---

📝 Citation

If you use this dataset, please cite:

@misc{misraj2025msdd,
  title        = {Misraj Structured Data Dump (MSDD)},
  author       = {Khalil Hennara, Muhammad Hreden, Mohamed Motaism Hamed, Zeina Aldallal, Sara Chrouf, Safwan AlModhayan, Ahmed Bustati},
  year         = {2025},
  publisher    = {MisrajAI},
  howpublished = {\url{[https://huggingface.co/datasets/Misraj/msdd](https://huggingface.co/datasets/Misraj/msdd)}}
}

---