---
language:
- ar
- ary
multilinguality:
- multilingual
size_categories:
- 10K<n<100K
task_categories:
- text-generation
- text-classification
pretty_name: darija-combined-dataset
tags:
- darija
- moroccan-arabic
- arabic
- north-african
---
# darija-combined-dataset
## Dataset Description
This dataset contains Darija (Moroccan Arabic) text data collected from multiple sources and processed for machine learning tasks.

### Dataset Summary
- **Total texts**: 35,562
- **Total characters**: 5,891,009
- **Average text length**: 165.7 characters
- **Number of unique datasets**: 8
- **Created**: 2025-07-28T15:03:41.001214
- **Processing time**: 18.93 seconds

### Source Distribution
- **Darija Pattern Dataset**: 1,076 texts
- **Audio Transcriptions**: 5,108 texts
- **wekipedia darija**: 4,862 texts
- **DODA Audio Dataset**: 10,246 texts
- **Filtered Samples Transcriptions**: 1,763 texts
- **IADD darija**: 5,590 texts
- **Openai Gsm8k Dataset**: 1,319 texts
- **Darija Banking Dataset**: 5,598 texts

### Dataset Composition
- **Darija Pattern Dataset**: 1,084 texts, 36,286 characters (avg: 33.5)
- **Audio Transcriptions**: 5,108 texts, 164,433 characters (avg: 32.2)
- **wekipedia darija**: 4,862 texts, 4,485,713 characters (avg: 922.6)
- **DODA Audio Dataset**: 10,504 texts, 265,018 characters (avg: 25.2)
- **Filtered Samples Transcriptions**: 1,765 texts, 161,988 characters (avg: 91.8)
- **IADD darija**: 6,355 texts, 225,997 characters (avg: 35.6)
- **Openai Gsm8k Dataset**: 1,319 texts, 297,053 characters (avg: 225.2)
- **Darija Banking Dataset**: 5,713 texts, 289,427 characters (avg: 50.7)

### Data Fields
- `text`: The text content in Darija
- `source`: The source dataset name

### Data Processing
The dataset was processed with the following configuration:
- **Minimum text length**: 10 characters
- **Maximum text length**: No limit
- **Duplicates removed**: Yes

### Usage
```python
from datasets import load_dataset

# Load the dataset
dataset = load_dataset("mohamed-stifi/darija-combined-dataset")

# Access the data
for example in dataset['train']:
    print(f"Text: {example['text']}")
    print(f"Source: {example['source']}")
```

### Languages
This dataset primarily contains:
- **Darija** (Moroccan Arabic) - ary
- **Modern Standard Arabic** - ar

### Citation
If you use this dataset, please cite:
```bibtex
@dataset{darija_dataset,
    title={Darija Dataset},
    author={Darija LLM Lab},
    year={2024},
    url={https://huggingface.co/datasets/mohamed-stifi/darija-combined-dataset}
}
```

### License
Please refer to the individual source datasets for their respective licenses.