Steel is the most widely used material among all metals. As the concern about effective use of the in-use steel stock as secondary resource grows, numerous studies analyzing the flow and stock of steel have been conducted. Though top-down and bottom-up approaches have been employed in material flow analysis, the applicability of these approaches is largely restricted by the data availability. To overcome this problem, we have proposed a method for estimating in-use steel stock based on satellite images. Our previous studies had shown effective regression of steel stock in buildings and nighttime lights on country and sub-national level. However, the relationship has not been examined in micro level, i.e. building stocks and pixels. Therefore, in this study, geographic information system (GIS) was used to estimate the in-use steel stock for buildings in Tokyo, Japan. The volume of buildings in 145 areas with 1 km square in Tokyo was calculated with GIS. Then, the correlation was investigated between the building volume, in-use building steel stock and the radiance of nighttime light. As the result, building volume and in-use building steel stock were found to be highly correlated with the radiance of nighttime light.