Data visualization is a powerful tool to cope with the demands of our current information age. In order to understand and be able to develop visualizations for specific use cases, data visualization activities (vis activities) have been proposed in recent years. These highly effective tools focus on practical relevance, reflection, and discussion in order to teach data visualization knowledge in a variety of contexts. However, the conscious selection of one or more vis activities for learners in comprehensive courses remains difficult. We aim to support this process by proposing a didactic vis framework. Based on Bloom's revised learning taxonomy, we decompose vis activities into distinct learning activities with their specific learning goals. By assigning the learning goals to the cognitive process and knowledge dimensions, a didactic course structure can be planned and evaluated. To demonstrate this didactic vis framework, we conducted several workshops based on an existing interface construction kit.