code-workout.Rmd

---
title: "Principle Component Aanalysis"
output: html_document
---

## Iin this repository is a data set and codebook from Rod Martin, Patricia Puhlik-Doris, Gwen Larsen, Jeanette Gray, Kelly Weir at the University of Western Ontario about people's sense of humor. 

```{r}
library(dplyr)
library(tidyr)

#Upload data and separate questions from composites
H1 <- read.csv("humor_data.csv", header = TRUE)
H2 <- select(H1, 1:32)

#Run pca on the questions
pca <- prcomp(H2, scale = TRUE)

summary(pca)

#Sum of Squared Distances/Eigenvalues

plot(pca$sdev)

# Loadings
pca$rotation

#Look at biplot, see how the questions resolve into four clear directions
biplot(pca)

#Look at the loadings, and compare them to the question text in the humor_codebook.txt, do they align with the composites that the authors generated (affiliative, selfenhancing, aggressive, self-defeating)?

pca$rotation

#They match closely, this is because the survey was designed to resolve into these four components.

```