In this paper we illustrate the Geo Data Annotator (GDA), a framework which helps a user to build a ground-truth dataset, starting from two separate geographical datasets. GDA exploits two kinds of indices to ease the task of man- ual annotation: geographical-based and string-based. GDA provides also a mechanism to evaluate the quality of the built ground-truth dataset. This is achieved through a col- laborative platform, which allows many users to work to the same project. The quality evaluation is based on annotator agreement, which exploits the Fleiss’ kappa statistic. 

Stefano Cresci

Davide Gazzè

Angelica Lo Duca

Andrea Marchetti

Maurizio Tesconi