DEVELOPMENT OF 3D MONITORING DATASET FOR 
SHOP AND OFFICE TENANTS VARIATIONS IN BROAD URBAN AREA BY 
SPATIO-TEMPORAL INTEGRATING DIGITAL MAPS AND YELLOW PAGE DATA 
Yuki Akiyama 3, *, Takeshi Shibuki b , Satoshi Ueyama 3 , Ryosuke Shibasaki c 
a Graduate School of Frontier Science, The University of Tokyo, Japan - aki@iis.u-tokyo.ac.jp, uym@csis.u-tokyo.ac.jp 
department of Civil Engineering, The University of Tokyo, Japan - tshibuki@iis.u-tokyo.ac.jp 
c Center for Spatial Information Science, The University of Tokyo, Japan - shiba@csis.u-tokyo.ac.jp 
KEY WORDS: Spatial Database, Spatio-temporal Data Mode, Image Understanding, Spatial-temporal Analysis, Urban Planning, 
Digital Mapping 
ABSTRACT: 
Various methods of analysis and dataset have been used for urban studies. However, used urban data often has low spatial resolution, 
especially regional statistics, though statistical data can cover large areas with homogeneous quality. For detailed analysis, some 
studies rely on field surveys which have fine spatial resolution, but fail to cover entire urban areas or large regions including suburbs 
and rural areas. Many prior studies also indicated the need of detailed time-series urban dataset all over Japan. In this study detailed 
time-series urban dataset in broad area have been developed. This dataset have been developed to integrate multi-year digital house 
maps and yellow page data using GIS system and developed programs in this study. Developed dataset cover South Kanto region for 
the period from 1995 to 2005 every five years. South Kanto region is the geopolitically important area of Japan including Tokyo 
metropolitan area.Using this dataset, not only name, business category, and detailed location including three-dimensional 
information but also time-series variation of almost all shops and offices can be monitored. Time-series variation means continuation, 
change, emergence, demise and immigration from other room or floor of tenants between two times. 
1. INTRODUCTION 
1.1 Background 
So far many kinds of dataset were developed for understanding 
“urban space”. However, development of dataset with high 
spatial resolution and reliability needs a large amount of labour. 
Therefore, they are able to cover limited area. On the other 
hand, spatial resolution and reliability of dataset it can cover 
large area is not so fine. Local statistical information is one of 
the examples of them. 
Since the collapse of the asset-inflated economy in Japan (after 
1991), wide range of urban problems have happened, for 
example, urban decay at many local cities. Time-series urban 
dataset which can closely monitor changes of urban space has 
been demanded to analyze and solve such problems. Previous 
studies also pointed out this necessity (Yasuyuki, M., 1994). 
In contrast, many skyscrapers are built in major large cities and 
as a result large cities increase its density of houses, shops and 
offices. Therefore, it seems that data which can monitor urban 
spaces in three dimensions is needed. Moreover, property 
investments expand in energetic urban areas in recent years. 
Due to the expansion of property investments, pace of urban 
change is increasing now as never before. 
To make a more detailed urban analysis, a method is required 
that can monitor spatial information including three 
dimensional distribution as micro as possible. A development 
of dataset is demanded that can monitor time-series variations 
of almost all shop and office tenants in broad area or all of 
urban areas, prefectures and national land. 
To realize such dataset, a development of methodology is 
needed that integrates appropriate existing multi-year spatial 
information at low cost and short time. 
1.2 Previous studies 
Most studies have also tried to develop spatio-temporal urban 
dataset using existing information. 
Itai developed 3D digital map around the Kawaguchi station 
(Saitama prefecture, Japan) that visualizes foundry industrial 
decline and urbanization (Youichi, I., 1994). This dataset was 
developed using distributional information of foundries in 1974 
and 1994. 3D time-series variation of foundries can be 
monitored using this dataset. However, base maps used in this 
study made manually. Therefore this method is not considered 
proper to develop dataset that can cover broad area. 
Ito developed a methodology to integrate digital house maps 
between different two years (Kaori, I., 2001). The methodology 
used in her study is helpful to us in conducting our research. 
However, this methodology was limited to integrate digital 
house maps. Moreover, dataset developed in her study is 
covered only the special words of Tokyo. Due to this limitation, 
it is difficult to apply this methodology in widely area than it. 
In contrast, some studies have collected time-series information 
using the questionnaire survey to local governments without 
relying on existing spatial information (Junichiro, A., 2002). 
Yuki Akiyama (M.Env.). Cw-503, Institute of Industrial Science, 4-6-1, Komaba, Meguro-ku, Tokyo, 153-8505, Japan, Phone: 
+81-3-5452-6417, Fax: +81-3-5452-6414, e-mail: aki@iis.u-tokyo.ac.jp