GIS Blog: Melanie A. Sutton

Striped burrfish are inshore fish known to exploit a wide range of depths. Although commonly found in grass beds around barrier islands in the Gulf of Mexico (Franks et al., 1972), they are also found near coral reefs (Robins, Ray, & Douglas, 1986), as well as in deeper waters during the winter months (Audubon Society, 2002). Changes to the geographic ranges of fish species due to factors such as ocean warming (Espino et al., 2019) are of concern to scientists. In addition, impacts to fish habitats in tourism-based states such as Florida are considered by economists given the socioeconomic benefits of commercial and recreational fishing to coastal communities (Lellis-Dibble, McGlynn, & Bigford, 2008). As one example of their impact, half of all federally managed fisheries in the U.S. depend on coral reefs (Burton, 2019). In addition, t he total value of coral reefs for southeast Florida was estimated in 2007 to be $174 million/year (Brander & van Beukering, 2013),

The map above demonstrates the use of bivariate choropleth mapping to demonstrate patterns in the United States related to obesity and physical inactivity, as highlighted in the close-up of the legend provided here:  In order to prepare data for bivariate choropleth mapping, two related variables must first be normalized. In this example, the variables have been normalized using the population of each county.  Next, each variable must be independently classed into an appropriate number of bins. The quantile classification method was selected in the example above so that the data values for each of the variables were distributed in equal numbers across the data bins - this meant there were no bins with too few or too many data points. Three classes were selected for each variable in this example to intuitively align with low, medium, and high rates of of physical inactivity in a population, as well as low, medium, and high rates for obesity in a population. Three class

This map uses proportional symbology to map positive and negative values. In order to achieve this within ArcGIS Pro, separate features classes representing positive ranges of values were created for the losses and for the gains.  When creating the final symbology for the two items, settings were then adjusted to make sure values noted in the legend were represented by similarly sized item on the map (i.e., same sized circle for 50K for jobs lost or jobs gained).  Finally, contrasting colors (green and orange) were then selected to help the reader visually compare states with losses and gains. A lighter border for the circles compared to the deeper shade in the interior was then used so the smallest sized items could be more easily discerned against the underlying layer with state borders, as shown in a close-up of the legend here:

Strategies employed when designing the final layout for this infographic included the following: Visual contrast: Lighter shades of blue and tan are used for areas with textual information to increase contrast with the black, blue, orange, and green fonts used in these areas. Additionally, the number of classes used for each choropleth map was selected to maximize the visual contrast between those counties with smaller percentages compared to those with large percentages. Legibility: The use of larger, bolded fonts for the “Don’t delay – get evaluated TODAY!” and “7 out of 10 diabetics have sleep disorders” textboxes was made to make these messages easily seen as the reader is scanning other supporting text provided in smaller or non-bolded fonts.  Figure-ground organization: The border around the center bar chart was thickened to make this visualization more prominent, as it demonstrates how one state (Alabama) has the three highest ranked counties for the summarized s

For this map, I created a customized version of intervals after starting out with an Equal Intervals classification. I used 9 classes to capture one class for negligible changes ( ± 0-0.01%) and four classes on either side of this. I selected two bin ranges for small population changes ( ± 0-2.5% and ± 2.5-5%), one bin size for medium population changes ( ± 5-10%), and one bin size for large population changes ( ± 10-20%). The selection of four classes on either side of negligible changes was done to more clearly represent the larger range of values for growth (versus shrinkage) in population trends. I used a divergent color scheme from ColorBrewer for the legend. I used a neutral gray color for the negligible change category and then selected 4 shades of green for positive change and 4 shades of red for negative change. The colors are shaded similarly for similarly-sized histogram bins on both sides of the range, and the darkest colors in both ranges are used to indicate the

Comparison of Color Ramps The three sequential, single-hue color ramps shown above demonstrate different methods for making a good selection of six contrasting colors.  In a linear progression (see A), the same interval within each of the RGB ranges is used for all steps. A disadvantage of this simple method is the contrast in the darker range is more difficult to discern.  One possible correction to this is to create an adjusted progression color ramp (see B), where the interval varies within each of the RGB ranges so that larger steps occur near the darker range.  Finally, use of the ColorBrewer tool (see C) is a faster alternative to creating a sequential, 6-color color ramp. In this case, the tool adjusts the steps within each RGB range to more optimally distinguish the colors at both the darker and lighter ends of the range.