emanfatimaa05/code-switching-codesaviours-si26-eman

Code Switching NLP Dataset

Code Saviours SI-26 | Week 6

Description

This dataset contains Roman Urdu-English code-switching sentences collected for Code Saviours SI-26 Week 6.

The purpose of this dataset is to provide examples of how Pakistani users naturally mix Roman Urdu and English in informal digital communication.

Dataset Contents

The dataset contains:

• 157 unique sentences
• 1,593 word-level entries
• 1,035 URD labels
• 558 ENG labels

Each word is labeled as either:

• URD — Roman Urdu / Urdu word
• ENG — English word

Data Collection

The sentences were collected from multiple sources, including:

• Pakistani social media content
• WhatsApp conversations, with personal information excluded
• Public online discussions
• Additional example sentences created to represent natural Roman Urdu-English code-switching patterns

The dataset focuses on informal language commonly used in Pakistani online communication.

Format

The dataset is provided as a CSV file with the following columns:

Column	Description
sentence	The complete code-switched sentence
word	Individual word from the sentence
label	Language label: URD or ENG

Example

sentence	word	label
Aaj ka din bohot busy tha	Aaj	URD
Aaj ka din bohot busy tha	busy	ENG
Aaj ka din bohot busy tha	tha	URD

Intended Use

This dataset can be used for experimentation with:

• Code-switching NLP
• Language identification
• Roman Urdu NLP
• Word-level language classification
• NLP models for Pakistani digital communication

Limitations

This is a relatively small dataset containing 157 sentences. It is intended primarily for educational and experimental purposes and should not be considered a comprehensive representation of all Roman Urdu-English communication.

License

This dataset is provided for educational and research purposes as part of the Code Saviours SI-26 Week 6 project.