A Bangla Language Movie Review Dataset with Sentiment and Emotion Labels

Published: 29 September 2025| Version 1 | DOI: 10.17632/48jjjrjvsn.1
Contributor:
Safaet Zahangir

Description

This dataset contains 1,500 unique Bangla movie reviews, carefully collected and designed to support research in Sentiment Analysis and Emotion Detection. Each review is written in natural Bangla, reflecting the way people usually express their opinions on movies. To make the dataset more realistic, the reviews vary in length and tone — some are short one-line opinions, while others are longer and combine multiple expressions, just like real comments found on YouTube or social media. Each review has been annotated with: Sentiment labels: Positive, Negative, or Neutral Emotion labels: Happy, Sad, Angry, Excited, Disappointed, or Neutral The dataset was built with special care to maintain logical consistency. For example, a review that expresses excitement about a movie is not labeled as Negative, and a clearly disappointed review is not marked as Positive. This makes the resource more reliable for training and evaluating natural language processing (NLP) models.

Files

Institutions

  • Daffodil International University

Categories

Natural Language Processing, Machine Learning, Emotion, Bangladesh, Human Language Resources, Sentiment Analysis

Licence