By improving the reliability and the durability of sensors as well as power supply technology and social infrastructure, all areas from medical and health care to transportation and agriculture are now covered by sensors. Thus a huge amount of data is available to be analyzed. The above situation is fueling the acceleration of the "Internet of Things (IoT)". At the Machine Learning and Data Science Center, we launched new research plan and developed spatio-temporal multi-dimensional collective data analysis that can predict when, where, and what in real time and provide feedback to social systems. In this talk, I introduce the concept and technology of our new approach.