Workshop20170817.Rmd

---
title: "Heatmaps"
author: "Joyce Robbins"
date: "8/17/2017"
output: html_document
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE, cache = TRUE, 
                      warning = FALSE, message = FALSE)
```


```{r, echo = FALSE}
library(ggplot2)
library(gridExtra)
```

Download this file from www.github.com/jtr13/RLadies

### geom_tile with numerical data, compare to geom_point  
```{r}
x <- 1:3
y <- c(5, 2, 7)
df <- data.frame(x, y)
g1 <- ggplot(df, aes(x, y)) + geom_point()
g2 <- ggplot(df, aes(x, y)) + geom_tile()
grid.arrange(g1, g2, nrow = 1)
```

### geom_tile with variable w, h
```{r}
df$w <- c(1.4, .4, .2)
df$h <- c(.5, 1.3, .8)
ggplot(df, aes(x, y)) + geom_tile(aes(width = w, height = h))
```

### geom_rect uses (xmin, xmax, ymin, ymax) instead
```{r}
xmin <- 1:3
xmax <- 2:4
ymin <- c(5, 2, 7)
ymax <- c(6, 3, 8)
df <- data.frame(xmin, xmax, ymin, ymax)
ggplot(df, aes(xmin = xmin, xmax = xmax, ymin = ymin,
               ymax = ymax)) + geom_rect()


```

### force squares
```{r}
ggplot(df, aes(xmin = xmin, xmax = xmax, ymin = ymin,
               ymax = ymax)) + geom_rect() + coord_fixed()
```

### geom_raster -- same as geom_tile w/ uniform w, h & FASTER
```{r}
x <- 1:3
y <- c(5, 2, 7)
df <- data.frame(x, y)

ggplot(df, aes(x,y)) + geom_raster()
```

### change x & y: what happens to the size of the tiles?
```{r}
x <- 1:2
y <- c(5, 2)
df <- data.frame(x, y)
ggplot(df, aes(x, y)) + geom_raster()
```

### change x & y again
```{r}
x <- 1:4
y <- c(5, 2, 10, 7)
df <- data.frame(x, y)
ggplot(df, aes(x, y)) + geom_raster()
```

# PRACTICE using geom_tile(), geom_rect(), geom_raster()

## Building the heatmap

### complete set of (x, y) pairs
```{r}
x <- c(1, 1, 1, 2, 2, 2, 3, 3, 3)
y <- c(1, 2, 3, 1, 2, 3, 1, 2, 3)
df <- data.frame(x, y)
ggplot(df, aes(x, y)) + geom_raster()
```

### add color
```{r}
set.seed(2017)
df$z <- sample(9)
ggplot(df, aes(x, y)) + geom_raster(aes(fill = z))
```

### What if z is categorical?
```{r}
df$z <- c("A", "B", "C", "B", "A", "A", "B", "C", "B")
ggplot(df, aes(x, y)) + geom_raster(aes(fill = z))
```

### What is z is discrete (numerical)?
```{r}
df$z <- sample(3, 9, replace = TRUE)
ggplot(df, aes(x, y)) + geom_raster(aes(fill = z))
df$z <- factor(df$z)
ggplot(df, aes(x, y)) + geom_raster(aes(fill = z))

```

### clean it up
```{r}
theme_heat <- theme_classic() +
  theme(axis.line = element_blank(),
        axis.ticks = element_blank())

ggplot(df, aes(x, y)) + geom_raster(aes(fill = z)) + 
  theme_heat
```

### add coord_fixed
```{r}
ggplot(df, aes(x, y)) + geom_raster(aes(fill = z)) + 
  coord_fixed() + theme_heat
```

### add white border (doesn't work with geom_raster)  

```{r}
ggplot(df, aes(x, y)) + 
  geom_tile(aes(fill = z), color = "white") + 
  coord_fixed() + theme_heat
```

### x & y are categorical
```{r}
grade <- rep(c("first", "second", "third"), 3)
subject <- rep(c("math", "reading", "gym"), each = 3)
statescore <- sample(50, 9) + 50
df <- data.frame(grade, subject, statescore)

ggplot(df, aes(grade, subject, fill = statescore)) + 
  geom_tile(color = "white") +
  coord_equal() + theme_heat
```

### order of categories
```{r}
grade <- rep(c("first", "second", "third", "fourth"), 3)
subject <- rep(c("math", "reading", "gym"), each = 4)
statescore <- sample(50, 12) + 50
df <- data.frame(grade, subject, statescore)

ggplot(df, aes(grade, subject, fill = statescore)) + 
  geom_tile(color = "white") +
  coord_equal() + theme_heat
# oops
```

```{r}
df$grade <- factor(df$grade,
                   levels = c("first", "second",
                              "third", "fourth"))
df$subject <- factor(df$subject,
                     levels = c("math", "reading",
                                "gym"))

ggplot(df, aes(grade, subject, fill = statescore)) + 
  geom_tile(color = "white") +
  coord_equal() + theme_heat
```

### Are there other ways to display 3 variables? Of course.
```{r}
ggplot(df, aes(x = grade, y = statescore, fill = subject)) +
  geom_col(position = "dodge")
```

# PRACTICE with Yamaguchi dataset from vcdExtra package
(or www.github.com/jtr13/RLadies -- click on "Yamaguchi87.rda", View Raw.  It should download, then click the file to open it.)  

Compare Japan, the UK, and the US in terms of occupational mobility.  

```{r}
library(vcdExtra)
library(dplyr)
orderedclasses <- c("Farm", "LoM", "UpM", "LoNM", "UpNM")
mydata <- Yamaguchi87
mydata$Son <- factor(mydata$Son, levels = orderedclasses)
mydata$Father <- factor(mydata$Father,
                        levels = orderedclasses)
japan <- mydata %>% filter(Country == "Japan")
uk <- mydata %>% filter(Country == "UK")
us <- mydata %>% filter(Country == "US")
```

### Japan

```{r}
ggplot(japan, aes(x = Father, y = Son)) + 
  geom_tile(aes(fill = Freq), color = "white") + 
  coord_fixed() + theme_heat
```

### UK
```{r}
ggplot(uk, aes(x = Father, y = Son)) + 
  geom_tile(aes(fill = Freq), color = "white") + 
  coord_fixed() + theme_heat
```

### US
```{r}
ggplot(us, aes(x = Father, y = Son)) + 
  geom_tile(aes(fill = Freq), color = "white") + 
  coord_fixed() + theme_heat
```

### All
```{r}
ggplot(mydata, aes(x = Father, y = Son)) +
  geom_tile(aes(fill = Freq), color = "white") +
  coord_fixed() + facet_wrap(~Country) + theme_heat
```

### All, as % of country total
```{r}
mydata2 <- mydata %>% group_by(Country) %>% 
  mutate(Total = sum(Freq)) %>% ungroup()

ggplot(mydata2, aes(x = Father, y = Son)) +
  geom_tile(aes(fill = Freq/Total), color = "white") +
  coord_fixed() + facet_wrap(~Country) + theme_heat
```

### All, as % of country x class total
```{r}
mydata3 <- mydata %>% group_by(Country, Father) %>% 
  mutate(Total = sum(Freq)) %>% ungroup()

g <- ggplot(mydata3, aes(x = Father, y = Son)) +
  geom_tile(aes(fill = Freq/Total), color = "white") +
  coord_fixed() + facet_wrap(~Country) + theme_heat
g
```

```{r}
g + geom_text(aes(label = round(Freq/Total, 1)),
              color = "white")
```

### Color
```{r}
library(pgmm)
data(wine)
library(tibble)
library(tidyr)
library(scales)
mywine <- wine %>% 
  rownames_to_column("id") %>% 
  gather(key, value, -id) %>% 
  mutate(key = factor(key)) %>% group_by(key) %>% 
  mutate(scaledvalue = rescale(value)) %>% ungroup()

ZtoAorder <- sort(levels(mywine$key), decreasing = TRUE)

mywine$key <- factor(mywine$key, levels = ZtoAorder)

ggplot(mywine, aes(x = id, y = key)) + 
  geom_raster(aes(fill = value)) + scale_fill_viridis()

ggplot(mywine, aes(x = id, y = key)) + 
  geom_raster(aes(fill = scaledvalue)) + 
  scale_fill_viridis()
```

### Take out type
```{r}
mywine <- wine %>% 
  rownames_to_column("id") %>% 
  gather(key, value, -id, -Type) %>% 
  mutate(key = factor(key)) %>% group_by(key) %>% 
  mutate(scaledvalue = rescale(value)) %>% ungroup()

ZtoAorder <- sort(levels(mywine$key), decreasing = TRUE)

mywine$key <- factor(mywine$key, levels = ZtoAorder)

ggplot(mywine, aes(x = id, y = key)) + 
  geom_raster(aes(fill = scaledvalue)) + 
  scale_fill_viridis() +
  facet_wrap(~Type, scales = "free_x")

ggplot(mywine, aes(x = id, y = key)) + 
  geom_raster(aes(fill = scaledvalue)) + 
  scale_fill_distiller(palette = "Purples") +
  facet_wrap(~Type, scales = "free_x")
```

## Yamaguchi87
```{r}
library(viridis)
grid.arrange(g, g + scale_fill_viridis(),
             nrow = 2)
```

```{r}
g + scale_fill_viridis(direction = -1)
```

```{r}
g + scale_fill_viridis(end = .9)
```

### RColorBrewer
```{r}
g + scale_fill_distiller(palette = "RdBu")
```

```{r}
ggplot(mydata3, aes(x = Father, y = Son)) +
  geom_tile(aes(fill = (Freq/Total)), color = "white") +
  coord_fixed() + 
  scale_fill_gradient2(low = "black", mid = "white",
                        high = "red", midpoint = .2) +
  facet_wrap(~Country) + theme_heat
```