Spatial Perspective for Indonesia Demography

We modify original queen based spatial weight matrix manually. Our result of spatial weight matrix is based on the combination of the simple (queen based) contiguity and the k-nearest neighbor based contiguity. Local Indicator of Spatial Autocorrelation (LISA) displays clusters and outliers. Clusters are shown in red (High-High value) and blue (Low-Low value). Other values (High-Low and Low-High) are respectively shown in light red and light blue. Regional demographic indicators are spatially autocorrelated significant at p=0.05. Distribution of each variable is by nature clustered rather than dispersed or random. Districts with relatively high value demographic indicators tend to be located near other relatively high value ones. District with relatively low value demographic indicators in each variables tend to be located near other relatively low values ones. Significant location of rural variable clusters are in western Java province, eastern Java province, eastern Nusa Tenggara province and south-eastern Sulawesi province. Significant locations of life expectancy hot spot are in Java island, while cold spots are located in Papua island.

Key words: spatial weight matrix, spatial data analysis, clusters