500,000 Indonesian news articles totaling 136 million words from seven major outlets including Detik, Kompas, and CNN Indonesia. The collection spans January 1, 2018, to August 20, 2018, and is provided in both raw JSON and cleaned text formats.
Use Cases
- Train Indonesian-specific word embeddings or transformers using the 136 million words in the newspapers.txt file
- Analyze media bias or reporting styles across the 7 different newspaper sources
- Extract events and entities from the 2018 Indonesian news cycle using the article text and date metadata
Strengths
- 500,000 articles from 7 Indonesian newspapers including Tempo, Republika, and Poskota
- 136 million words of Indonesian text data for language modeling
- Temporal coverage spanning January 1, 2018, to August 20, 2018
- 2.2GB of uncompressed data provided in JSON and cleaned text formats