Purpose – To design a ‘small multiples’ layout that shows four choropleth maps, depicting spatial distribution of the data side by side with different results to depict different classification methods. For sake of comparison, each map was designed the same way with the same aggregated attribute data, and number of classes and colour scheme; using the Goodness of Variance Fit (GVF) to determine the validity of classification methods.

 

Method – The base map data was downloaded in shapefile format and the attribute data in CSV format. Four data frames were then created for each classification method, with an additional header and footer.

 

The projected coordinate of State Plane Pennsylvania, South, FIPS 3702 NAD 1983 in meters was used. The first layout was created using CMYK neutral colour values of 8, 8, 16, 0 for the data frame background and 12, 12, 24, 5 for the header and footer fill.

 

The objective of data classification is to group data into classes where the values in each class are as like each other as possible, and the classes themselves are as different from each other as possible. The data classification method with the highest assessment index value should be used.

 

The data was spit into five classes, using the formula GVF = (SDAM-SDCM)/SDAM. Where SDAM was the sum of the squared deviations from the array mean and the SDCM was the sum of the squared deviations from the class mean. Any GVF value above 0.7 is considered to be a high GVF.

​

Graduated colours, quantitative map type was chosen, with a field ratio value of White Population 2010. Equal Interval classification was chosen with five classes. Each class fill and outline colour was then manually changed to distinguish the colour on the map with the corresponding colour on the legend. The colours were set using the “Colour Brewer” chart by Cynthia Brewer, from Penn State University.

 

Next, using the Legend Wizard, the legend was ungrouped and manually modified to Arial, size 8 labels, Arial Bold, size 9 for the legend title. Each box was then resized to 20 points width and 15 points in height with a vertical patch gap of zero, so there was zero space between the boxes. The legend title, boxes and continuous class ranges were then grouped together and carefully aligned with the edges of the data frame to 0.1” between the left and bottom edges of the frame.

 

The Equal Interval data frame was then copied and pasted into the top right data frame and Quantile classification method is chosen. And the copied legend inherited the properties of the Quantile method. The procedures were the same for the Natural Breaks classification. The Manual classification map used upper limit classes of 3.0, 6.0, 12.0, 24.0 and 48.0.

​