ant_test.Rmd

---
title: "ant_nest_nutrients"
author: "Allison Horst"
date: "March 6, 2018"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

Short PCA example (commented out so you can see what happens one step at a time...)

###Get packages and ant nest nutrient data
```{r}
library(corrplot)
library(tidyverse)

ant_nutrients <- read_csv("ant_nutrients.csv")
ant_df <- rename(ant_nutrients, Na = X1) # Rename because it thinks it's NA
```

```{r}
corrplot(cor(ant_df), method = "ellipse")
# Other option:
corrplot(cor(ant_df), type = "upper", method = "square")


# Perform PCA

ant_pca <- prcomp(ant_df, scale = TRUE) # Note: all must be numeric (the country names are just labels), so we exclude column 1
summary(ant_pca) # Within first two PCs, 63% of variance explained!

# Use the plot() function on the PCA name to see how the variances decline as the principal component increases

plot(ant_pca)
# Or, alternatively:
screeplot(ant_pca, type = "lines") # Shows the same information

# Notice that a large amount of the variance is described after 3 - 4 principal components (but we just show two on the biplot)

# Check out the biplot! 

biplot(ant_pca, cex = 0.5)
```