realKD

Creedo is a web application that provides an intuitive graphical user interface for realKD algorithms and allows to rapidly design, deploy, and conduct user studies. Its target audience are data mining researchers who want to showcase their algorithms through an easily accessible demonstrator and/or to empirically evaluate their innovations with real users. Creedo’s source code is available from Bitbucket under the MIT license and there are also binary releases, which just require Java 8 and MySQL Server (5.5+) to run.

Analytics Dashboards

Creedo analytics dashboards provide an intuitive user interface for supporting analysts in an iterative knowledge discovery process. In every round, the user can run a pattern discovery algorithm, investigate the results within the context of the data, store those results that she finds suitable to her current task, and then re-iterate. The exact content of a dashboard can be configured by the dashboard designer depending on its purpose and target audience.

Creedo Studies

In order to test how well certain analytics dashboards support users to perform specific data analysis tasks, Creedo allows to define designs for empirical user studies. Studies based on these designs can then be executed by Creedo any number of times as long as a suitable pool of study participants is available. As a further essential feature, Creedo studies can not only involve users in the role of test participants, but also in the role of evaluators who evaluate study results based on customized critera that require human judgement. The application automatically performs all task assignments to participants and evaluators, who can then carry out their tasks using any standard web browser.

A peer reviewed research paper on Creedo was published at IDEA 2015.

Also see paper on One-Click Mining for the description of an intelligent interactive data analysis system running through Creedo.