A series of earthquake swarms off the east coast of Izu Peninsula, central Japan, is typically associated with dike intrusion around active volcanoes. Numerous studies in this region have documented the seismic activity and ground deformations without fully resolving the process of dike intrusion. This study involves a joint analysis of geodetic and seismic data from the most recent major activity in this region, occurring in 1998, to reveal the process of dike intrusion quantitatively. We combine data from a dense GPS array, including 10 temporary stations, with seismic analysis published previously. We estimate the location of the dikes from seismic data and the volume of intruded magma from geodetic data. Combined analysis of both data sets enables us to reveal the kinematic process of dike intrusion associated with the 1998 earthquake swarm. The process of dike intrusion is quantitatively modeled using the time evolution of recorded dike volume and the use of seismic hypocenters as an indicator of the site of instantaneous dike intrusion. On the basis of these observations, we propose the following process of the dike intrusion: The magma rose upward through a pathway aligned vertically and created a dike when it lost buoyancy. We also evaluate the dike intrusion process quantitatively using a model of a circular dike that grows under a continuous magma supply from a deep magma reservoir. From the observational data, we infer the time evolution in the dike shape, pressure distribution, and magma viscosity in the dike. The acquired viscosity increases over time and it may be reasonable because magma cools gradually. However, its value is too large compared to typical basaltic magma expected in this region. It may represent magma within the dike that is nearly solidified. Otherwise, the magma is supplied intermittently and the dike growth is mainly controlled by the magma supply rate from a deep reservoir. The quantitative description of the dike intrusion process in nature described in this study is very rare, and our results provide constraints on several parameters that have not been solved in many previous theoretical studies on the dike intrusion process.