It's exciting times to work with data: more and more data are becoming available; more people are getting trained in analysis; better analysis methods provide us with more and better information about our world. With this came more and trickier problems which require better solutions for analysis and to supporting analysts. In this talk, I present examples and interactive interfaces of how data visualization supports analysts along a range of tasks, including data collection, exploration (http://vistorian.net), annotation, understanding machine learning (http://playground.tensorflow.org), and explaining insights findings (http://datacomics.net). These examples are giving an overview over how to apply data visualization and which problems visualization research is dealing with. The overarching question I'd like to discuss is how can we provide better access and training for the wide variety of visualization techniques, tools, and interfaces?