Loading datasets...
Release Date: 7/15/2026
Format: CSV, XLSX
Size: 252.75 KB
Share
The Sindhi Sentiment Analysis Dataset is an open NLP resource containing 4,420 manually labeled sentences in the Sindhi language, annotated for three sentiment classes: positive, negative, and neutral. Every sentence has been hand-labeled to ensure linguistic accuracy and cultural relevance. Sindhi is a low-resource language spoken by over 30 million people in Pakistan and India, yet it remains severely underrepresented in NLP research and digital language tools. This dataset addresses that gap by providing a structured, high-quality text corpus for sentiment analysis and text classification tasks. The dataset is also available on Kaggle (https://www.kaggle.com/datasets/alinawaz06/sindhi-sentiment-analysis-dataset) and HuggingFace (https://huggingface.co/datasets/alinawazmahar/Sindhi_Sentiment_dataset). It is intended to serve researchers, linguists, and developers working on Sindhi language technology and low-resource NLP.
Licensing
Creative Commons Attribution Non Commercial 4.0 International (CC-BY-NC-4.0)
https://spdx.org/licenses/CC-BY-NC-4.0.htmlRestrictions/Special Constraints
This dataset is intended for research and scientific purposes only. Users must provide proper attribution to the original author (Ali Nawaz, Shaikh Ayaz University) when using or redistributing this dataset.
Forbidden Usage
This dataset may not be used for commercial purposes without prior permission. Users may not misrepresent the dataset, its source, or its labels.
Ethical Review
The dataset consists entirely of text sentences with no personal, sensitive, or identifying information. All data was collected and labeled by the author with full awareness of its intended research use. No human participants were involved in data collection, and no ethical concerns regarding privacy or consent apply to this dataset.
Intended Use
This dataset is intended for training and evaluating sentiment analysis models for the Sindhi language. Example applications include opinion mining, social media sentiment monitoring, customer feedback analysis, and building NLP benchmarks for low-resource South Asian languages.
| Column | Description |
|---|---|
sentence | Sindhi text sentence |
sentiment | Label: positive, negative, or neutral |
Total sentences: 4,420
Positive: ~1,500
Negative: ~1,500
Neutral: ~1,420
Language: Sindhi (sd)
Script: Perso-Arabic
sindhi_sentiment_v2.csv — primary dataset
sindhi_sentiment_cleaned.xlsx — cleaned version
Sentences were collected from diverse Sindhi text sources covering social, cultural, and everyday topics. All sentences were manually labeled by the author. The dataset was expanded using back-translation and synthetic generation, with all additional sentences verified and labeled manually.
| Column | Description |
|---|---|
sentence | Sindhi text sentence |
sentiment | Label: positive, negative, or neutral |
Total sentences: 4,420
Positive: ~1,500
Negative: ~1,500
Neutral: ~1,420
Language: Sindhi (sd)
Script: Perso-Arabic
sindhi_sentiment_v2.csv — primary dataset
sindhi_sentiment_cleaned.xlsx — cleaned version
Sentences were collected from diverse Sindhi text sources covering social, cultural, and everyday topics. All sentences were manually labeled by the author. The dataset was expanded using back-translation and synthetic generation, with all additional sentences verified and labeled manually.
Suitable for training and evaluating sentiment classifiers, text classification benchmarks, and low-resource NLP research for Sindhi.
