In recent years

international migration has become more and more common

while its patterns are getting increasingly complex.After several years living in a certain destination

immigrants often make repeat migrations

either to return to their home country or onward to another host country.Researchers have developed new methodologies to study the flows and patterns of transnational movements

and new theories to explain them.One of the new developments in this respect is the triangular model of human capital transfer among nations by Devoretz and Ma（2002）

which emphasizes the dynamic nature of international migration in the context of regional development.It argues that immigrants often enter into an entrepot country to accumulate transnational human capital and other capital

such as citizenship.After acquiring such capital

they can choose to stay in the country

return to their home countries

or move to another host country

in order to maximize the return to their acquired transnational capital.New challenges are posed to the study of transnational flows of human capital

requiring the use of multiple censuses of the sending country

entrepot country and major hosting country.Due to structural differences

censuses from different countries are normally in different format and cover populations within different national boundaries.As a result

the most current research on international migration is often limited to either the sending or receiving countries.Efforts are needed to integrate these data sets into a standardized one

so that variables can be unified for direct comparison and further analysis.We adopt a new approach to the integration of census micro-data of different countries into one unified framework.The framework includes two major parts: the statistical part and the mapping part.We start with the statistical part by using open source software packages

such as PHP and MySQL

to implement an integrated micro database system.The micro database includes censuses from multiple countries/regions

including the US

Canada and Hong Kong.In order to enable automated analysis

we first select common variables from different censuses and then standardize each of them to the same unit or category.These standardized variables are either called identifiers or indicators.Identifiers are variables used to identify similar population groups from different censuses and indicators are variables used to compare among groups.In the demo system

we used a total of 10 identifiers and 3 indicators.With the integrated database

we designed a search module and a statistics module.The search module uses key identifiers to search specific population groups from different censuses.The result is listed as tabulations to support further studies.The statistics module takes previous tabulations as input and output results of statistical analysis

including cross country/region comparison and uni-variable analysis（maximum/minimum/mean/std） as tables

graphs

and pre-map files.The second part of the framework is to integrate the micro database with a GIS database

the result of which can be used for mapping purposes.The statistical outputs from Part 1 are processed by the mapping module of Part2

and the system creates online map visualization.This paper mainly introduces the first part of the framework.We are still working on the second part.When both parts are finished and integrated

this system will integrate spatial information with census micro-data of different countries/regions

and provide a unified web-based demographic information system to facilitate flexible and advanced international immigration analysis.