Omar-youssef/QA_Finance_Egyptian_dataset

QA Finance Egyptian Dataset

A synthetic Arabic (Egyptian dialect) question-answering dataset focused on financial and accounting topics, such as financial statement analysis, corporate performance evaluation, and related concepts.

Dataset Details

• Language: Arabic (Egyptian dialect, ar)
• Domain: Finance / Accounting
• Size: 2,670 examples (train split)
• Format: Question–answer pairs, each tagged with its source topic

Dataset Structure

Each row contains:

Column	Type	Description
question	string	A finance-related question written in Egyptian Arabic
answer	string	The corresponding answer, also in Egyptian Arabic
source_topics	string	The topic(s) the QA pair was generated from

Example

{
  "question": "إيه الهدف الأساسي من تحليل القوائم المالية للشركات؟",
  "answer": "الهدف الأساسي هو تقييم الأداء المالي للشركة بشكل شامل، تحديد نقاط قوتها وضعفها، وتقدير قدرتها على تحقيق الأرباح وسداد التزاماتها في المستقبل، وده بيساعد في اتخاذ قرارات استثمارية أو إدارية صحيحة.",
  "source_topics": "تحليل القوائم المالية للشركات"
}

Dataset Creation

The QA pairs were synthetically generated from a curated set of finance-related topics, then structured into question/answer form in Egyptian Arabic.

Usage

from datasets import load_dataset

dataset = load_dataset("Omar-youssef/QA_Finance_Egyptian_dataset")
print(dataset["train"][0])

Intended Use

This dataset is intended for training and evaluating Arabic (Egyptian dialect) language models on finance-domain question answering, instruction tuning, and retrieval-augmented generation tasks.