As color and geometry representations are easily recognized by human's brain, data visualization technologies provide data mining results with natural and intuitive operation interfaces. Currently, data mining users always plot 2D and 3D diagrams using the traditional data visualization tools. However, it is difficult to visualize big data in real time using these tools. To enhance the visualization quality, this paper proposes a real-time data mining and visualization system for big data. We apply a graphic processing unit (GPU) programming technology to enhance the data mining algorithms and visualization methods in parallel. For unified development environment supports, we utilize a TCP/IP to deliver datasets between different data mining platforms and databases. In the visualization client, this system provides two human computer interaction methods, including a graphics user interface (GUI) and a natural user interface (NUI). To verify the efficacy of our proposed system, we integrated it with an NUI engine, the WinSocket API and the DirectX.