Finding Patterns in Census Data Using ArcGIS Geostatistical Analyst

Task One: Compare the First Four Patterns

The Swiss Federal Statistical Office publishes choropleth maps of census data on approximately 3,000 communities. Patterns are not always easy to discover in maps of the whole country, and it is even harder to compare two or more maps. Interpolations can be used to extract (e.g., simplify) these patterns, and if they are done using kriging, the amount of information retained can be measured.

Solution: Use ArcGIS Geostatistical Analyst to Simplify Patterns First

Task Two: Compare the Following Four Patterns

The two maps on the lower right show very similar patterns: tertiary graduates and graduates of upper secondary general education live in similar places (they prefer the shores of lovely lakes and can afford them). People with only compulsory education (second row, center) live in rural and industrial areas. Those with a vocational upper secondary education form a third pattern quite distinct from the other two. The percentage of information retained ranges between one-half (highest and lowest education) and one-third of the variance (correlations predicted—measured values are: r = .73, .76, .69, .63).