Browse Datasets

Tunizi

Tunisian Arabizi Sentiment Analysis

sentiment-analysis text-classification ~3,000 sentences MIT
Access Dataset Read Paper

Description

First sentiment analysis dataset for Tunisian Arabizi (Tunisian dialect written in Latin script with numerals).

Quick Start

Python 
from datasets import load_dataset

dataset = load_dataset("chaymafourati/tunizi", split="train")

Project Ideas

Arabizi Tokenization

Handle Arabizi (numbers as letters) in tokenization

Intermediate

Dialect Comparison

Compare MSA-trained model vs Tunizi-trained model

Intermediate

Ethical Considerations

  • Social media data privacy considerations