Urbanisation and spatial inequalities in health in Brazil and城市化和空间的不平等健康巴西.pptVIP

* These other index, in the evenness/clustering dimension, we are not going to use in our work, so, lets move on. 7s * It was for this dimension that a new index was developed. From those available from the software, one, the Neighbourhood Sorting Index was only global. And without the local indexes it was useless. And the Spatial Dissimilarity Index was not linear, it was high both in the richest and the poorest areas, and could not be used in the regression analysis. So, to measure the segregation in the evenness/clustering dimension we developed a new measure based on the Local Indices of Autocorrelation. * At the level of the census tracts, The Moran Scatter plot and the Moran Cluster map were produced. Each census tract was classified into four categories. In blue are depicted the census tracts which mean income of the head of the households was below the global mean and were surrounded by other census tracts also with income below the global mean. In red are the census tracts with high income and that are surrounded by other high income census tracts. The others are the outliers: in light blue are the low income tracts that are surrounded by high income tracts, and in pink, are the high income tracts surrounded by low income census tracts. This quadrant represents the low-low situation. Low income census tracts (on the left in the X axe) surrounded by low income neighbours, below, in the Y axe. As each district has some 30 census tracts, for each district we calculated the proportion of low income census tracts that are surrounded by other low income census tracts. 2min * And here are our final results for this new Spatial Clustering Index. In brown are the districts where 100% of its the census tracts have income below the global mean and are surrounded by other low income census tracts. In light yellow are the districts where none of its census tracts were low income tracts surrounded by other low income tracts. We think that for the