Proportional Symbols & Bivariate Maps

 

I admittedly could not think of a good way to superimpose two maps onto one another like the instructions of the lab suggested while maintaining the same proportional symbol sizes.  What I did instead was use one map and set the US_States_Albers layer to proportional symbology.  For the field selection, I used a custom selection of Abs($feature.JOBS) to set the proportions of the symbols by the absolute value of the Jobs field.  Next, in the Vary Symbology by Attribute section of the Symbology Pane, I set the field to Jobs, set the color scheme type to a Discrete Color Scheme, and only used two colors (red for negative values, green for positive).  I set the red color to stop at a value of 0 on its histogram in the pane and set the green color value to begin at the first positive value (1,200 jobs created in Delaware).  I exported the US boundary from the countries shapefile from the last map and placed it under the layer driving the proportional symbols; because of the proportional symbols, there was no color fill in the states.  I added the Job Loss and Job Gain layers that were created in the previous steps to the legend, put them as the last layer so they wouldn’t render in the layout, and gave them the same HEX# as their respective symbols from the US_States_Albers layer.

Preparing data and symbology for bivariate maps is more meticulous and time consuming than it is difficult.  Because bivariate maps aren't rendered in the ArcGIS world, manually manipulating a data set is the first step.  I first created three fields in the data, one for each variable to be mapped and one that will combine the two variables essentially.  Once the variables are identified, using a three-class quantile classification method to determine the class breaks in the data is necessary in order to populate the newly added fields.  Using the Calculate Geometry and querying out a selection based on the variables class breaks, two of the fields are populated relatively easily, albeit manually.  Lastly, combing the two sequential fields (1,2,3 and A,B,C) across a row of data will provide you with nine classifications that can be used in a bivariate map (1A, 1B, 3C, etc.).

Finding logical symbology is a bit tricky at first.  By starting out with two opposite color schemes, colorbrewer2.org is a great resource to get started.  It can be tricky to adjust the hue, saturation, and value of four symbols that aren't part of a three-class sequential color scheme that can be easily found online.  Incrementally adjusting HSV values to find the intermediate colors was laborious but necessary.  The resulting product is a map that shows correlation between two variables using choropleth symbology and one map.