Visual Analytics is the integration of interactive visualization with analysis techniques to help answer questions, and find interesting"patterns" in a given dataset. This approach is useful in cases where human knowledge and intuition is required. Further, with the increasing size of data, an interactive visual interface can give the user control over what she wants to see. Towards this end, we develop a Visual Analytics (VA) infrastructure, rooted on techniques in machine learning and logic-based deductive reasoning. This system assists people in making sense of large, complex data sets by facilitating the generation and validation of models representing relationships in the data. The above data is often communicated among people, and by systems as graphs and other visual forms. However, currently the degree of semantics encoded in these representations is quite limited. In order to design more expressive icons, we present a framework that uses natural language text processing and global image databases to help users identify metaphors suitable to visually encode abstract semantic concepts. While using images as forms of communication, people often like to add annotations in forms of text, arrows, etc. to make them more expressive. The process of selecting a color for the annotation can often be difficult if the background has a variation in color and texture. Our tool Magic marker helps users by suggesting appropriate colors for annotating an image. All colors in a modified CIE-Lab color-space are assigned a preference value, and users can navigate this space using an interactive interface. This interactivity also allows them to make the final choice based on personal preferences.