A Large Scale Tweet Dataset for Urdu Text Sentiment Analysis

Description

This is a dataset of tweets in the Urdu language. There are 1,140,824 tweets in the dataset, collected from Twitter for September and October 2020. This large-scale corpus of tweets is generated by performing preprocessing, which includes removing columns containing user information, retweet’s count, followers, and duplicate tweets, removing unnecessary punctuations, links, symbols, and spaces, and finally extracting emojis if present in the tweet text. This dataset's final tweet record contains columns for tweet id, text, and emoji extracted from the text with a sentiment score. Emojis are extracted to validate Machine Learning models used for the multilingual sentiment and behavior analysis.

Institutions

Categories

Licence