Packages and Data

The Data

 The data used for both plots (referred to as 'bee') was obtained from a study on Common Eastern Bumble Bees. The study tested the relationship between the size of the bee workers and the success of the colony. One indiator of colony success is reproduction, which was measured by the number of gynes (female bees with reproductive capabilites) in the colony. Gynes were distinguished from worker bees by their size: above 9 mm was labeled a gyne. Bee size was determined by thorax width, measured at the widest point and to the closest millimeter (mm).      
 The location of the hives was also documented in the data. The 62 colonies used in the study were located within 1.375 ha patches in 3 varying shapes: connected, rectangle, and winged. The patches were surveyed for the richness of plant species, and were labelled 2 to 8, with 8 being the highest amount of plant species, and 2 being the least. 
 The second data set, referred to as 'bee_sum' is a summary of the original data, but manipulated to pick out the mean thorax width of the workers and gynes, and to include the count of each.

Bee Size and Gyne Count

#R> `geom_smooth()` using method = 'loess' and formula 'y ~ x'

This first plot “Bee Size and Gyne Count,” explores one of the main questions from the study: how does bee size affect the reproductive success of the colony? I used the summarized data ‘bee_sum’ becuase I needed to single out the mean thorax width of worker bees and the gyne count. Those variables were not distinguished in the original data set.
I modeled by plot after a scatterplot published in the study’s paper. Having a scatterplot to explore this relationship appears to be the clearest way to convey the results. The points are slighty transparent, which I did to help see the overlapping points that occur along the bottom of the plot. I decided to use a LOESS line to show the general trend. I changed the color to red, which I feel drew more attention to it, and adjusted the alpha to make the confidence band lighter. I considered removing the axis titles but I wanted the units of measurements for both variables to be clear. From the scatterplot, there appears to be a slightly significant relationship between worker bee size and gyne count, with the gyne count generally increasing as the mean thorax width increases.Although there is some variation in the plot, the LOESS line shows the general trend of increasing mean thorax width and gyne count.

Plant Richness and Bees Size Among Hive Patch Shape

In this next plot, “Plant Richness and Bee Size Among Hive Patch Shape,” I wanted to explore the effect that plant richness had on the thorax width of bees. With a basic scatterplot, there doesn’t appear to be a significant relationship between the two, as the thorax width varies greatly among the different levels of plant richness. Next, I wanted to see if the shape of the patch where the hive was located had any significance on either variable. By faceting the plot by the ‘Patchtype’ variable, I could compare thorax width and plant richness for each patch shape. From that, I observed that the winged patch shape had the highest plant richness. The rectangle patch shape had plant diversity in each level from 2 to 6, while the connected patch had more variety, with values from 3 to 7. The graph doesn’t explain why this difference exists, which would probably be a different study altogether. For my design choices, I chose a scatterplot to allow patterns among the x- and y-axis to be clear, and adding geom_jitter showed the density of these points. I changed the titles of the x- and y-axis to better explain the variables. I placed limits and breaks on the y-axis to keep the breaks within the data range and to allow the viewer to see the values without the axis being over-crowded. For my color choice, I used a color-blind friendly palette, with soft colors that don’t distract the eyes from the overall plot. I faceted using facet_wrap() becuase I only wanted to facet with one variable, which was patch shape. To make the facet labels stick out a little more, I changed the lettering to bold. I added a title, subtitle, and caption and made adjustments to the theme by removing the legend from faceting, and removing the gridlines on the x-axis. Like the first plot, I thought about removing the axis labels, but decided not to because I wanted the viewer to understand that plant richness was on a scale of 2-8, and that bee size in measured by thorax width in millimeters.