Our third analysis was done to highlight areas with similar income level, number of schools, and park area to see what the relationship was between these three qualities.
To do the grouping analysis, we had to put the data on median income, number of schools, and park area in each census tract together in one layer. First, we added a field for “count” to the school layer and placed a value of 1 for each school. Then, we joined the school layer to the census tract layer and summed the values so it totalled the number of schools within the CT. To find the park area within the CT, we used the same method as in our buffer analysis.
The first grouping analysis we did used four groups, but one of the groups was thrown off because Rouge Park in the northeast corner of Toronto takes up such a large portion of the CT that it was put into its own group. Because this census tract is such a significant outlier, we decided to exclude it so it would not throw off the entire analysis. We removed that CT from the layer, and did the grouping analysis again. We tried multiple group amounts and chose five as the best amount because it was the maximum number of groups where each group was still unique from the others.