Week3_Homework.R

library(tidyverse)

wings <- read_csv("PureLines_Clean.csv")

wings$Line <- factor(wings$Line)
wings$Real_Sex <- factor(wings$Real_Sex)

## JD: Any reason you're not being "tidy"? ie., using mutate and eschewing "$"?

#for this week I want to visulize the relationship between size and shape in my data.
#first, to be able to visulize shape data, I need to do a PC decomposition


## JD, there is no wings_alt. I'll do a link, but you have to tell me if that's what you meant
wings_alt <- wings
PCs <- prcomp(wings_alt[, 6:101])
 
#adding PC values to the wings data frame 
#We keep the first 56 PCs (number of demetions in data) explains ~99.95% of varriance 
 
summary(PCs)

wings_PCs <- data.frame(wings_alt, PCs$x[,1:56]) 

#Now I will plot the wing size vs PC1. Because there is a scaling step during shape data collection, no PCs will nessisarity be assosiated with size
#Assosiation between size and shape indicates an allometric effect on wing shape 
#The vectors are going to be really short because we don't typically see a large varriation in wing size within a line 
## JD: Not clear what you mean by this, but you could get more resolution here by saying scale="free" (or something) in your facet_wrap.

#Plotting real data with a trendline
#if anything, these realationships look hypoallometric 
## JD: Can you explain "hypoallometric"?
## JD: What is "real" about Real_Sex? Should it be called sex?

ggplot(wings_PCs, aes(x = log(CS), y = PC1, color = Real_Sex)) + 
  geom_point(alpha = 0.2, size = 0.5) +
  geom_smooth() +
  facet_wrap(~Line)

## JD: You really only did one thing, plus a facet_wrap.
## Grade: good (2/3)