BengaliNewspaperCommonCoverageArticleDataset
Description
The dataset has been gathered from five well-known newspapers in Bangladesh. It contains 1056 titles and under each title (per row), the dataset provides five sets of information corresponding to each newspaper, including the date, URL, headline, and description. Data Format: • Articles are encoded into UTF-8 text file. • The title is in the first row of each file. • The content of the article starts from the second-row forwards. Each row represents one article with a unique serial number (SL), along with its common Bangla headline and publication date (yy/mm/dd format). For each article, the dataset includes the article URL links and the corresponding headlines and descriptions from five major Bangladeshi newspapers: Prothom Alo Daily Inqilab Daily Ittefaq Bangladesh Pratidin Kaler Kantho This allows comparison of how the same news article was reported, titled, and described across different newspapers.
Files
Steps to reproduce
The dataset was generated by manually searching and comparing articles across five Bangladeshi e-newspaper portals. Articles were included only when a common headline was found across all sources. Article URLs, headlines and descriptions were extracted using web browsing, Avro Keyboard for Bangla typing, and a Chrome Image-to-Text Scanner extension when articles appeared only as images. Data were cleaned, formatted and compiled using Excel/Numbers, ensuring consistent structure and reproducibility.
Institutions
- American International University Bangladesh