Twitter dataset is a collection of Bangla tweets downloaded by querying Twitter REST API v1.1 over a span from May-November 2013. As Twitter API supports language filtering and allows specifying the language of the retrieved posts, the optional language parameter in the Twitter Search URL was set to ‘bn’ to extract all Bangla tweets. Twitter datasets has two polarities - positive, negative.

Blog Datasets are the collections of Bangla comments from different bangla blogs. Blog datasets has six polarities- sad, happy, disgust,surprise,fear, angry.