Tweet Conversation Annotation Tool with a Focus on an Arabic Dialect, Moroccan Darija

This paper presents the DATOOL, a graphical tool for annotating conversations consisting of short messages (i.e., tweets), and the results we obtain in using it to annotate tweets for Darija, an historically unwritten Arabic dialect spoken by millions but not taught in schools and lacking standardization and linguistic resources. With the DATOOL, a native-Darija speaker annotated hundreds of mixedlanguage and mixed-script conversations at approximately 250 tweets per hour. The resulting corpus was used in developing and evaluating Arabic dialect classifiers described briefly herein. The DATOOL supports downstream discourse analysis of tweeted “conversations” by mapping extracted relations such as, who tweets to whom in which language, into graph markup formats for analysis in network visualization tools.