Advances in data acquisition and storage technology lead to the creation of increasingly large, complex data sets across many different application domains, including science, engineering, business and social media. Often, this data is of complex nature, involving high-dimensional and temporal data. Important user tasks for leveraging large, complex data sets include finding relevant information, exploring for patterns and insights, and re-using of data for authoring purposes. If appropriately combined, methods from interactive data visualization and data analysis allow to factor in background user knowledge during the analysis process, and provide solutions for many search and analysis tasks in important applications. In this talk, we discuss visual-interactive data analysis techniques from our work that can support search and analysis in a variety of different data types and novel application scenarios. These include example- and sketch-based retrieval in scientific data repositories, interactive cluster analysis approaches for time-oriented data, and user-adaptive learning of patterns in high-dimensional data. We also touch upon the problem of analysis of local patterns in global data sets, by example of local scatter plot patterns. We conclude with an outline of research challenges in the area.