Context Natural Language Process, or NLP, is one of the most studied machine learning field. Much progress has been made in recent years which has allowed this field to move into large scale use in several domains. Nowadays, NLP is widely used in social networks, in search engines, in translation tools, in chatbot assistants, and many others cases ... However, the progress and results differ from language to another. The majority of machine learning models treat in priority English and abandon the other languages, in particular Arabic. The main reason for this is the lack of datasets. Hence all the interest of the current datasets gathering +5000 news articles in Arabic. Content This data-set contains 5870 news articles in Arabic language extraced from aljazeera.net website. Each item contains the following information : - guid - published - title - description - link - content - image - ref - tags Files aljazeera.db - an sqlite database file article_aljazeera.csv - a export of data in csv format article_aljazeera.json - a export of data in json format Inspiration NLP is a large fields of machine learning, that try to resolve many issues related to language. Here is some of this issues for inspiration : - a summarization task - an embedding model (like Word2vec of Glove) - a classification model to identify article domain (ex : propose a list of tags based on the article) - a sentiment classification - a co-reference identification model - a named-entity recognition - ...