farunawebservices/Igala-Dataset-Explorer-App

igala-streamlit-app # 🌍 Igala Dataset Explorer Interactive Streamlit application for exploring and auditing low-resource language datasets. Built for the first comprehensive Igala-English parallel corpus with 3,253 field-collected sentences. ## 🎯 Overview This tool provides **dataset quality insights** for NLP researchers working with low-resource languages: - Sentence length distributions - Vocabulary coverage analysis - Translation pair alignment quality - Linguistic feature statistics - Data collection methodology documentation ## 🚀 Live Demo Explore the dataset: huggingface.co ## 📊 Dataset Statistics | Metric | Value | |--------|-------| | **Total Sentence Pairs** | 3,253 | | **Igala Vocabulary Size** | 8,742 unique tokens | | **English Vocabulary Size** | 6,318 unique tokens | | **Avg Sentence Length (Igala)** | 12.4 tokens | | **Avg Sentence Length (English)** | 11.8 tokens | | **Collection Method** | Field recordings + manual transcription | | **Quality Verification** | Human-verified by native speakers | ## ✨ Features - ✅ **Interactive Visualizations**: Word clouds, length distributions, token frequency plots - ✅ **Search & Filter**: Find specific sentence pairs by keyword or length - ✅ **Quality Metrics**: Alignment scores, vocabulary overlap, translation consistency - ✅ **Export Options**: Download filtered subsets in CSV/JSON - ✅ **Linguistic Insights**: Morphological complexity, tone marker frequency ## 🛠️ Tech Stack - **Frontend**: Streamlit - **Data Processing**: Pandas, NumPy - **Visualization**: Plotly, Matplotlib, WordCloud - **Deployment**: HuggingFace Spaces ## 📦 Installation ```bash # Clone the repository git clone github.com cd igala-dataset-explorer # Install dependencies pip install -r requirements.txt # Run the app streamlit run app.py 🔍 Usage Load and Explore Dataset import pandas as pd from dataset_utils import DatasetAnalyzer # Load dataset df …