Datasets:

cazzz307
/

malayalam-kannada-tamil-telugu-samam-dataset

Modalities:

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

xet

Community

Access Request for Malayalam Books Dataset

It may take maximum 24hours to process your request.

To access this dataset, please provide the following information. You agree to use this dataset for non-commercial purposes only and not to conduct experiments that cause harm to human subjects. These brief questions help us ensure access goes to genuine researchers or developers with clear intent—making our approval process more secure and fair.

license: cc-by-4.0 task_categories: - translation - text-classification language: - ml # Malayalam - kn # Kannada
- ta # Tamil - te # Telugu multilinguality: - multilingual size_categories: - 1K<n<10K source_datasets: - original tags: - dravidian-languages - multilingual-dictionary - malayalam-script - translation - linguistics - cross-lingual dataset_info: features: - name: page dtype: int64 - name: row
dtype: int64 - name: malayalam dtype: string - name: kannada dtype: string - name: tamil dtype: string - name: telugu dtype: string config_name: default splits: - name: train num_bytes: 2500000 num_examples: 4000 - name: validation num_bytes: 125000 num_examples: 200 download_size: 2625000 dataset_size: 2625000 configs: - config_name: default data_files: - split: train path: train.csv - split: validation
path: validation.csv default: true

Samam.net Multilingual Dictionary Dataset

Dataset Description

This dataset contains multilingual dictionary entries scraped from samam.net, a comprehensive South Indian language dictionary. The dataset provides translations between Malayalam and three other Dravidian languages: Kannada, Tamil, and Telugu.

Important Note about Script Usage

All text in this dataset is written in Malayalam script, even for non-Malayalam languages. This is a key characteristic of this dictionary resource:

Malayalam column: Malayalam words written in Malayalam script
Kannada column: Kannada translations of Malayalam words, but written in Malayalam script (not Kannada script)
Tamil column: Tamil translations of Malayalam words, but written in Malayalam script (not Tamil script)
Telugu column: Telugu translations of Malayalam words, but written in Malayalam script (not Telugu script)

This makes the dataset particularly valuable for Malayalam speakers who want to understand equivalent words in other South Indian languages while reading them in familiar Malayalam script.

Dataset Structure

The dataset contains the following columns:

page: Source page number from samam.net (1-493)
row: Row number within the page
malayalam: Malayalam word/phrase in Malayalam script
kannada: Kannada translation in Malayalam script
tamil: Tamil translation in Malayalam script
telugu: Telugu translation in Malayalam script

Dataset Statistics

Total entries: ~14,790 translation sets (30 entries per page × 493 pages)
Languages covered: 4 (Malayalam, Kannada, Tamil, Telugu)
Script: Malayalam script for all languages
Source pages: 493 pages from samam.net glossary

Use Cases

This dataset is valuable for:

Cross-linguistic research in Dravidian languages
Malayalam-centric translation systems
Comparative linguistics studies of South Indian languages
Educational resources for Malayalam speakers learning other Dravidian languages
Script conversion research (Malayalam script to native scripts)
Multilingual NLP applications focused on Dravidian languages

Data Quality

All entries have been cleaned to remove HTML artifacts
Language suffixes (ma, ka, ta, te) that appeared due to scraping have been removed
Empty or meaningless entries have been filtered out
Text has been normalized for consistent formatting

Unique Characteristics

Script Homogeneity: Unlike typical multilingual datasets where each language uses its native script, this dataset uses Malayalam script throughout
Malayalam-centric: Designed from the perspective of Malayalam speakers
Comprehensive Coverage: Covers everyday vocabulary across four major South Indian languages
Cultural Context: Reflects how multilingual dictionaries are traditionally organized in Kerala

Potential Applications

Research Applications

Study of phonetic similarities between Dravidian languages
Analysis of loanwords and linguistic influence
Comparative morphology research
Script standardization studies

NLP Applications

Training Malayalam-script multilingual models
Building cross-lingual embeddings for Dravidian languages
Developing translation systems for Malayalam speakers
Creating language learning applications

Educational Applications

Teaching aids for Dravidian language learning
Comparative vocabulary studies
Cultural and linguistic education resources

Limitations

Script Limitation: Non-Malayalam languages are not in their native scripts
Accuracy: As with any web-scraped data, there may be occasional errors
Coverage: Limited to vocabulary present in the source website
Transliteration: The Malayalam script representation of other languages may not capture all phonetic nuances

Data Collection

The data was collected through web scraping of the samam.net glossary using Python with the following approach:

Respectful scraping with delays between requests
Multi-threaded collection for efficiency
Error handling and retry mechanisms for reliability
Data validation and cleaning post-collection

Citation

If you use this dataset in your research, please cite:

Samam.net Multilingual Dictionary Dataset
Source: https://samam.net/glossary/
Collection Date: August 2025
Languages: Malayalam, Kannada, Tamil, Telugu (all in Malayalam script)

License

This dataset is compiled from publicly available web content. Please respect the original source's terms of use and consider this dataset for research and educational purposes.

Technical Details

Format: CSV, JSON, Excel
Encoding: UTF-8
Character Set: Malayalam Unicode range (U+0D00–U+0D7F)

Sample Entry

{
  "page": 1,
  "row": 1, 
  "malayalam": "അടുത്ത",
  "kannada": "ഹത്തിര",
  "tamil": "അടുത്ത",
  "telugu": "ദഗ്ഗരി"
}

Note: All four language entries above are written in Malayalam script, representing the word "near/close" in each respective language.

Contact

For questions about this dataset or issues with the data, please open an issue in the repository.

Downloads last month: 27