This technology automatically generates a phylogenetic tree from about 30,000 base sequence data of the new coronavirus collected worldwide. This visualizes the infection routes and mutation patterns. For instance, it illustrates how the virus spread from Wuhan and arrived in Japan through two initial routes; a direct route from China and another route via Europe.

To optimize the timing of road repairs, it is essential to understand the state of road deterioration. We introduce a method that can quantitatively and easily capture the state of road deterioration. Road degradation causes up-and-down movements in vehicles traveling on it. This up-and-down movement is detected by an acceleration sensor built into a commercially available dashcam, and the vibration data is transferred to a cloud server. In the cloud server, the vibration data is converted into frequency spectrum data, and by creating a two-dimensional map, the deteriorated state of the road section that caused the vibration can be color-coded and visualized. Finally, by assigning colors according to the state of deterioration for each road section on the road map, you can see at a glance the state of deterioration of each road section on the road map.

In face of situation where a part of the thousands of users at the computer center may have been unauthorized access, we used access log data to identify potentially malicious users. We mapped the access data of all users with D’Insight. Although the data of normal usage patterns formed the main cluster, some data were significantly distributed away from the main cluster, leading us to believe there might be unauthorized users involved.

A request came from a heart disease specialist to extract effective test elements for distinguishing between two types of heart failure (HFrEF and HFpEF) from the 44 test items for heart diseases. Using visual AI to evaluate effectiveness, we concluded that five elements were effective. By mapping the data of over 200 patients based on the test results of these five elements, we were able to cleanly separate the two types of heart failure patients, normal people, and error data, demonstrating the possibility of a new simplified diagnosis method for heart failure.