---
title: "Precept 5"
author: "Emily Nelson"
date: "March 2, 2016"
output: html_document
---

#Fun With `ggplot2` and Alcohol

```{r setup, message=FALSE}
library(dplyr)
library(ggplot2)

load("wine.RData")

wine <- wine %>%
  tbl_df()
```

#Boxplots / Violin Plots

What does a boxplot / violin plot show us?

1. Show a boxplot of the pH for each quality score. (Why do we have to cast?)

2. Add a `color` layer for wine type. What happens? What if you use `fill` instead?

3. Change the colors manually. (What if you wanted to change the `fill` colors manually?)

4. Spruce up the axis labels and title.

5. Change the `theme`.

6. Change it to a violin plot. What is the difference?

```{r boxplots}

```

#Histograms

What does a histogram show us?

1. Make a histogram of the alcohol content of the wines with appropirate axis labels. (What happens if I give `geom_histogram` both an `x` and a `y` value?)

2. Add a `fill` layer based on type with manual color selection. What happens?

3. Add a vertical line showing the mean of the data. (How would you add a horizontal line?)


```{r histograms}

```

#Barplots

What does a barplot show us?

1. Make a barplot showing the mean quality scores for high and low alcohol content wines. Use appropriate themes, labels, etc.

2. Add a color layer. What happens? Is this a good way to visualize the information?

3. Make it better.

```{r barplots}
wine %>%
  mutate(alcohol.amt = ifelse(alcohol > mean(wine$alcohol), "high", "low")) %>%
  group_by(type, alcohol.amt) %>%
  summarize(mean.quality = mean(quality)) 
```


#Scatter Plots

What does a scatter plot show us?

1. Make a scatter plot showing the relationship between the pH and the fixed acidity, using an appropriate theme and labels.

2. Add a trend line using `stat_smooth`. Try both `loess` and `lm` types. How are they different?

3. Add a color layer for the type of wine. What does this do? What happened to the trend line?

4. Make the points transparent since there is a lot of overlap.


```{r scatterplots}
  
```

#Facets

The above plot still doesn't look great. There is too much overlap. We can fix this with facets!

1. Add a facet layer. What should it be?

2. Change the color of the trend line.

3. Free the x and y scales. What happens?

```{r facets}

```


#Hexbin

Change the above plot to a `hexbin` plot.

Note: may need to install the `hexbin` package.

```{r hexbin}

```