Abstract
Psychologists are interested in predicting the personality traits of individuals, which costs them great effort. The researchers in the field of computer science have found in the social media platforms an easier way achieve these studies. As they are widespread and used by most individuals. Social media enables people to share most of their thoughts, feelings and daily activities. Researches all-over the world tried to use that rich source of data to predict personality traits in many different languages, but none of them have tried to work with the Arabic language. Although Arab social media users represent a large sector of social media users. This paper introduces AraPersonality. A Personality traits dataset that was gathered from Egyptian dialect twitter users. It is consisted of about 92 users twitter feeds. This paper discusses how it was gathered, its properties, statistics and experiments, its experimental results can be considered the baseline. The average of the best result in each trait is 0.59 and 0.33 in binary and multiclass representation respectively according to f1-score. These results can be considered as a good start for a system that predict personality for Egyptian users.