The description of the process that transforms non graphic data (astronomical, biological, or financial data for example) into interactive graphical representations of this data has been at the core of information visualization research over the past decades. That process often referred to as coding, can be supported and controlled at different levels and by different types of users. Different levels bring different types of opportunity. In order to make it possible for all users to benefit from the pros of different levels approaches, we propose a toolkit named STOOG that addresses the needs of different types of users. STOOG brings a four-fold contribution: (1) a style sheet language for the description of the transformation process at stake in information visualization, (2) an application building automatically interactive visualization of data based on the style-sheet description, (3) a user interface devoted to the conception of style sheets and then (4) the generation of basic legends. This paper presents STOOG basic concepts and mechanisms and provides a case study to illustrate the benefits of using STOOG for the interactive and visual exploration of information.