It is vital for the transportation industry, which performs most of its work by automobiles, to reduce its accident rate. This paper proposes a 3D visual interaction method for exploring caution areas from large-scale vehicle recorder data. Our method provides (i) a flexible filtering interface for driving operations such as braking or handling operations by various combinations of their attribute values such as velocity and acceleration, and (ii) a 3D visual environment for spatio-temporal exploration of caution areas. The proposed method was able to extract caution areas where some accidents have actually occurred or that are on very narrow roads with bad visibility by using real data given by one of the biggest transportation companies in Japan.