AIAnastasia/Georgian-attractions

Georgian Tourist Attractions Dataset 🇬🇪

Dataset Description

This dataset contains bilingual (English and Russian) information about Georgian tourist attractions, accompanied by images. The dataset is specifically designed for educational and research purposes and is optimized for building RAG (Retrieval-Augmented Generation) systems, multimodal search applications, and tourism-related AI projects.

Attraction Categories

The dataset includes diverse types of Georgian attractions:

Architectural Heritage:

• Churches and cathedrals
• Historical bridges
• Monasteries and religious complexes
• Fortresses and ancient structures

Natural Attractions:

• Lakes and water bodies
• Rivers and gorges
• National parks and protected areas
• City parks and gardens
• Waterfalls and natural monuments

This comprehensive coverage makes the dataset ideal for training models that need to understand both cultural heritage and natural tourism aspects of Georgia.

Dataset Statistics

• Total records: 901
• Full bilingual records: 727
• Russian-only records: 174
• Images: 1,187
• Languages: English, Russian

Dataset Structure

Fields

• id: Unique identifier for each attraction
• name_en/ru: Name of the attraction in English/Russian
• description_en/ru: Detailed description in English/Russian
• location_en/ru: Location information in English/Russian
• category_en/ru: Category/type of attraction in English/Russian
• ner_en/ru: Named Entity Recognition data in English/Russian
• image_en/ru: Associated images for English/Russian context
• source: Indicates data source ("en", "ru", or "en, ru")

Intended Use

Educational and Research Purposes Only

This dataset is intended exclusively for educational and research purposes. It should be used for:

• Academic research and studies
• Model training and experimentation
• Educational projects and coursework
• Algorithm development and testing

Commercial use is not permitted without proper licensing.

Optimal for RAG Systems

This dataset is specifically optimized for RAG (Retrieval-Augmented Generation) applications:

• Rich contextual information: Detailed descriptions in two languages provide comprehensive context for retrieval
• Multimodal content: Combined text and images enable advanced multimodal RAG implementations
• Structured metadata: Categories, locations, and NER data facilitate precise retrieval
• Bilingual coverage: Enables cross-lingual RAG systems for international tourism applications
• Geographic specificity: Focused on Georgian attractions for domain-specific RAG performance

Usage Examples

Loading the Dataset

from datasets import load_dataset

# Load the full dataset
dataset = load_dataset("AiAnastasia/georgian-attractions")

# Access the data
example = dataset["train"][0]
print(f"Name (EN): {example['name_en']}")
print(f"Name (RU): {example['name_ru']}")

# Access images
image_en = example["image_en"]  # PIL Image object
image_ru = example["image_ru"]  # PIL Image object

Filtering by Language

# Only complete bilingual records
bilingual = dataset["train"].filter(lambda x: x["source"] == "en, ru")

# Only English records
english_only = dataset["train"].filter(lambda x: "en" in x["source"])

# Only Russian records
russian_only = dataset["train"].filter(lambda x: "ru" in x["source"])

For RAG Applications

# Example: Create embeddings for semantic search
from sentence_transformers import SentenceTransformer

model = SentenceTransformer('all-MiniLM-L6-v2')

# Combine text fields for embedding
def create_text(example):
    en_text = f"{example['name_en']} {example['description_en']} {example['location_en']}"
    ru_text = f"{example['name_ru']} {example['description_ru']} {example['location_ru']}"
    return en_text + " " + ru_text

# Create embeddings
embeddings = model.encode([create_text(ex) for ex in dataset["train"]])

Applications

RAG System Development

• Tourism RAG Chatbots: Build intelligent assistants that can retrieve and generate information about Georgian attractions
• Cross-lingual RAG: Develop systems that can answer questions in English or Russian using the bilingual content
• Multimodal RAG: Create systems that combine text retrieval with image understanding
• Domain-specific RAG: Train specialized retrieval systems for tourism and cultural heritage

Research Applications

• Multimodal Search: Combine text and image search capabilities for comprehensive attraction discovery
• Cross-lingual Information Retrieval: Study bilingual information retrieval techniques
• Tourism Recommendation Systems: Build personalized recommendation engines
• Cultural Heritage Preservation: Digital archiving and accessibility projects

Educational Projects

• NLP Coursework: Practice with real-world bilingual datasets
• Computer Vision: Work with culturally rich image datasets
• Information Systems: Build tourism information platforms
• Machine Learning: Train classification and recommendation models

Dataset Creation

This dataset was created by merging English and Russian CSV files containing attraction information with corresponding image files. The data was processed to ensure consistency and completeness while preserving all available records, including:

• 727 complete bilingual records with both English and Russian descriptions
• 174 additional Russian-only records providing extended coverage
• 1,187 high-quality images of Georgian attractions
• Comprehensive metadata including categories, locations, and NER annotations

License and Usage Terms

** Educational and Research Use Only**

This dataset is made available exclusively for educational and research purposes under the Creative Commons Attribution-NonCommercial 4.0 International License. By using this dataset, you agree to:

• Use the data only for non-commercial, educational, or research activities
• Properly cite the dataset in any publications or presentations
• Not redistribute or modify the dataset without permission
• Respect the cultural and intellectual property represented in the images and descriptions

For commercial licensing or other use cases, please contact the dataset creators.

Citation

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

@dataset{georgian_attractions_2024,
  title={Georgian Tourist Attractions Dataset},
  author={AiAnastasia},
  year={2024},
  url={https://huggingface.co/datasets/AiAnastasia/georgian-attractions},
  note={Bilingual dataset of Georgian tourist attractions with images for RAG applications}
}