An Arabic twitter data set of 7,503 tweets. The released data contains manual Sentiment Analysis annotations as well as automatically extracted features, saved in Comma Separated (CSV) and Attribute-Relation File Format (ARFF) file formats. Due to twitter privacy restrictions we replaced the orig...