9_Graphing_With_ggplot2.Rmd

---
title: '8) Combining and Reshaping Data'
author: "Jasmine Hughes"
date: "9/12/2020"
output: html_document
---


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


```{r}
# If you haven't yet installed dplyr, install it now!
# install.packages('dplyr')
library(dplyr)

# Lets also load the medals data set:
medals <- read.csv("data/all_medalists.csv", stringsAsFactors = FALSE) 
View(medals)
head(medals)
head(gap)
head
```

# Why Graph?

* Communicate your results to others.
    + Involved, fancy
    + May invest time in making it eye-catching, communicative
    + Seen by many people (colleagues, publications)
* Check calculations, assumptions, understand your data prior to analysis
    + Quick and dirty
    + Target audience is you
    + Get a quick answer to a question you have about the data, see if anything "smells wrong" about the data.

**Visualize your data often and early: a couple "quick and dirty" graphs can save you a lot of time.**

Graphing is often iterative, and you may make many graphs of the same data in an analysis:

* The first graph you make to check the data
* The graph you make that helps you get to that "Aha!" moment
* The graph you make that helps you communicate to others what the main finding is.

# How Graph?

3 main facilities for producing graphics in R: **base**, **`ggplot2`**, and **`lattice`**

When picking a graphing tool, consider:

* need for speed, vs need for polished final graph
* size of data (ex: very big data sets and many points, vs <5000 points)
* specialty graphs (beeswarm, ridge plots) versus basic graph types (bar, scatter)
* plays nicely with other "specialty" tools you are using.

# The Data

* All Summer Olympic medalists from 1896-2008
* Variables include name, gender, country, country code (`NOC`), sporting event, and type of medal won
* We don't actually care about names of winners: we're interested in how many medals that different countries won

For more info, see [The Guardian Datablog Olympic medal winners: every one since 1896 as open data](https://www.theguardian.com/sport/datablog/2012/jun/25/olympic-medal-winner-list-data)

First, we'll use `dplyr` and `tidyr` to count the medals of each type won, for each country and each year.

```{r}
head(medals)
# dplyr and tidyr refresher
medal_counts <- medals %>%
  group_by(Medal, Year, NOC) %>%
  summarise(count = n())
head(medal_counts)


```

This table is in tidy format. Wide (untidy) format can be useful for plotting in *base* plot (more on this later)

```{r}
medal_counts_wide <- medal_counts %>% 
  spread(key = Medal, value = count) %>%
  ungroup() %>%
  mutate(Bronze = ifelse(is.na(Bronze), 0, Bronze)) %>%
  mutate(Silver = ifelse(is.na(Silver), 0, Silver)) %>%
  mutate(Gold = ifelse(is.na(Gold), 0, Gold))
head(medal_counts_wide)
```

Finally, let's subset the data to gold medal counts for the US, for easier plotting.

```{r}
usa_gold_medals <-  medal_counts %>%
  filter(Medal == "Gold") %>%
  filter(NOC == "USA")
```

# Base graphics

We saw a couple examples of graphing in base R previously: the histogram and the boxplot.

```{r}

pt_data <- data.frame(MRN = rep(c(97, 107, 402),2),
                      measurement_time = factor(c(rep("Pre-Intervention", 3), 
                                           rep("Post-Intervention", 3)),
                                           levels = c("Pre-Intervention", "Post-Intervention")),
                      measurement_value = c(50, 65, 47, 104, 82, 91))


boxplot(measurement_value ~ measurement_time, data = pt_data)

hist(pt_data$measurement_value)
```

Base R is not super consistent about the graphing API (application programming interface - how you interact with a pieces of software). 

## Basic Base R plots

The general call for base plot looks something like this:

```{r, eval=FALSE}
plot(x = , y = , ...)
```
Additional parameters can be passed in to customize the plot:

* type: scatterplot? lines? etc
* main: a title
* xlab, ylab: x-axis and y-axis labels
* col: color, either a string with the color name or a vector of color names for each point

More layers can be added to the plot with additional calls to `lines`, `points`, `text`, etc.

```{r, fig.cap = ""}
plot(medal_counts_wide$Year, medal_counts_wide$Gold) # Basic
plot(usa_gold_medals$Year, usa_gold_medals$count, type = "l",
     main = "USA Gold Medals",
     xlab = "Year", ylab = "Count") # with updated parameters
```

## Other plot types in base graphics

These are just a few other types of plots you can make in base graphics.

```{r, fig.cap = ""}
boxplot(Gold~Year, data = medal_counts_wide)
hist(medal_counts_wide$Gold)
plot(density(medal_counts_wide$Gold))
barplot(usa_gold_medals$count, width = 4, names.arg = usa_gold_medals$Year,
                               main = "USA Gold Medals")
mosaicplot(Year~Medal, medal_counts)
```

*Aside*: many functions take a "formula" as an argument, in the format : "y ~ x1 + x2 + ....". We'll see this in more detail when we cover statistics in R.

## Object-oriented plots

Remember how `summary()` returns different answers depending on if the column is a factor, a character or a number? Base R can be "smart" about your data type too.

* Base graphics often recognizes the object type and will implement specific plot methods
* lattice and ggplot2 generally **don't** exhibit this sort of behavior

```{r, fig.cap=" "}
# Calls plotting method for class of the dataset ("data.frame")
plot(medal_counts_wide %>% select(-NOC))


# Calls plotting method for class of medal_lm object ("lm"), print first two plots only
medal_lm <- lm(Gold ~ Bronze + Silver, data = medal_counts_wide)
summary(medal_lm)

plot(medal_lm, which=1:5)
```

Can you tell R was made by statisticians?

# Pros/cons of base graphics, ggplot2, and lattice

Base graphics is

a) good for exploratory data analysis and sanity checks

b) a handy tool to use in tandem with `lm` and a few other 'base R' stats tools

c) inconsistent in syntax across functions: some take x,y while others take formulas

d) defaults plotting parameters are ugly and it can be difficult to customize

e) that said, one can do essentially anything in base graphics with some work

`ggplot2` is

a) generally more elegant

b) more syntactically logical (and therefore simpler, once you learn it)

c) better at grouping, handling multiple variables

d) has a well-developed ecosystem of packages that work well together, great for specialized graph types

e) the "quick and dirty" graph is often a reasonably aesthetically pleasing starting spot

`lattice` is

a) faster than ggplot2 (though only noticeable over many and large plots)

b) may be simpler than ggplot (although, fewer resources online for learning)

c) able to do 3d graphs

d) used by many specialty packages (bioinformatics, cytometry)

We'll focus on ggplot2 as it is very powerful, very widely-used and allows one to produce very nice-looking graphics without a lot of coding.

# ggplot2

The general call for `ggplot2` graphics looks something like this:

```{r, eval=FALSE}
# NOT run

ggplot(data = ) +
  geom_xxxx(aes(x = ,y = , [options]))) + 
  geom_xxxx(aes(x = ,y = , [options]))) + 
  ... + 
  ... + 
  ...
```

Note that `ggplot2` graphs in layers in a *continuing call* (hence the endless +...+...+...), which really makes the extra layer part of the call.

Careful: `+` is not the same as `%>%` (common typo!) We are adding layers together, not piping the input of one function into another.


You can see the layering effect by comparing the same graph with different colors for each layer:

```{r, fig.cap=" ", warning=FALSE}
p <- ggplot(data = medal_counts_wide) +
  aes(x = Year, y = Gold) +                   # unquoted column names: ggplot is part of the tidyverse
  geom_point(color = "gold")
p
p + geom_point(aes(x = Year, y = Silver), color = "gray") + ylab("Medals")
```

# Grammar of Graphics

`ggplot2` syntax is very different from base graphics and lattice. It's built on the **grammar of graphics**.
The basic idea is that the visualization of all data requires four items:

1) A dataset and set of mappings from variables to **aesthetics** (ex: year -> x, medal count ->y). In ggplot, use `aes()`

2) A **geometry** (`geom_xxx`), which describes how the aesthetics data is transformed into the graphical representation in terms of **kind** (or **shape**).

3) One **scale** for each aesthetic mapping used. Scales describe how the aesthetics data is transformed into the graphical representation in terms of **degree**. (What does a value of `5` or `10` when determining what each pixel in a graph looks like?). 

  * `scale_color_manual` and family for mapping values to colors/linetypes/circle types (displayed in a figure legend)
  * data transofmrations (log, etc), described using gridlines and axes for position scales

4) A coordinate system, affects all position variables (cartesian, logarithmic/semi-log, polar)


*Aside*: "The Grammar of Graphics" is a book, if you're interested in data visualization theory!




`ggplot2` allows the user to manipulate all four of these items.

Each layer is built up of the following components:

1) Aesthetic mapping (and data) (`ggplot(aes(x = Species, y = Sepal.Length))`)

2) A statistical transformation. (`geom_bar(stat = 'count')`, `geom_smooth(method = 'lm')`)

3) Geometric object (`geom_point()`, `geom_bar()`)

4) A position adjustment (ex: how to handle collision between points)


```{r, warning = FALSE, message = FALSE, fig.cap = "", fig.height=2.5}
ggplot(medal_counts_wide, aes(x = Year, y = Gold)) + 
  geom_point() +
  ggtitle("Gold Medal Counts")

ggplot(usa_gold_medals, aes(x = Year, y = count)) + 
  geom_line() +
  ggtitle("USA Gold Medals")

# Boxplots
ggplot(medal_counts_wide, aes(x = factor(Year), y = Gold)) +
  geom_boxplot() + 
  ggtitle("Gold Medal Counts")

# Histogram
ggplot(medal_counts_wide, aes(x = Gold)) + 
  geom_histogram() +
  ggtitle("Gold Medal Counts")

# Density plot
ggplot(medal_counts_wide, aes(x = Gold)) + 
  geom_density() +
  ggtitle("Gold Medal Counts")

# Bar chart
ggplot(usa_gold_medals, aes(x = Year, y = count)) + 
  geom_bar(stat = "identity")
```

* Consistent API call
* Change out one `geom` for another to change the "type"

# `ggplot2` and tidy data

* `ggplot2` plays nice with `dplyr` and pipes. If you want to manipulate your data specifically for one plot but not save the new dataset, you can call your `dplyr` chain and pipe it directly into a `ggplot` call.

```{r, fig.cap = "", fig.show = "hold"}
# This combines the subsetting and plotting into one step
medal_counts %>%
  filter(Medal == "Gold") %>%
  filter(NOC == "USA") %>%
  ggplot(aes(x = Year, y = count)) + 
  geom_line()
```

* Base graphics/lattice and `ggplot2` have one big difference: `ggplot2` **requires** your data to be in tidy format. For base graphics, it can actually be helpful *not* to have your data in tidy format.
The difference is that `ggplot` treats `Medal` as an aesthetic parameter that differentiates kinds of statistics, whereas base graphics treats each (year, medal) pair as a set of inputs to the plot.
Compare:

```{r, fig.cap = ""}
usa_all_medals <- medal_counts %>%
  filter(NOC == "USA")

# ggplot2 call
ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
            geom_line(aes(color = Medal))
```

```{r, fig.cap = ""}
usa_all_medals_untidy <- medal_counts_wide %>%
  filter(NOC == "USA")

# Base graphics call
plot(usa_all_medals_untidy$Year, usa_all_medals_untidy$Gold, col = "green",
                                 type = "l")
lines(usa_all_medals_untidy$Year, usa_all_medals_untidy$Silver, col = "blue")
lines(usa_all_medals_untidy$Year, usa_all_medals_untidy$Bronze, col = "red")
legend("top", legend = c("Gold", "Silver", "Bronze"),
                fill = c("green", "blue", "red"))
```


# Pros/cons of `ggplot2`

* Allows you to add features in "layers"
* Automatically adjusts spacing and sizing as you add more layers (iterative graphing)
* Requires data to be in tidy format
* Syntax is different from base R -- there is a learning curve
* Plots are actually objects. You can assign them to a variable and do things with it (more on this later)

# An overview of syntax for various `ggplot2` plots

## densities:

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}

ggplot(data = usa_gold_medals, aes(x = count)) + 
  geom_density() 
```

## X-Y scatter plots:

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point() 

ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_jitter() 
```

## X-Y line plots: 

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_line() # ggplot2
```

## bar plots:

```{r, fig.cap=" ", warning=FALSE, message=FALSE, fig.show = "hold"}

median_gold_medals <- medal_counts %>%
  filter(Medal == "Gold") %>%
  # mutate(count = as.double(count)) %>%
  group_by(NOC) %>%
  summarise(med = median(count))

head(median_gold_medals)


median_gold_medals %>%
  head() %>%
  ggplot() + 
  aes(x= NOC, y = med) +
  geom_bar(stat = 'identity')

 medal_counts %>%
  filter(Medal == "Gold") %>%
   ggplot() +
   aes(x = count) +
   geom_bar(stat = 'count')

```

## boxplots:

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
# Notice that here, you must explicitly convert numeric years to factors
ggplot(data = medal_counts_wide, aes(x = factor(Year), y = Gold)) +
            geom_boxplot() # ggplot2
```

## "trellis" plots:
```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
# Subset the data to North America countries for easier viewing
northern_hem <- medal_counts_wide %>%
  filter(NOC %in% c("USA", "CAN", "CUB", "MEX"))

ggplot(data = northern_hem, aes(x = Year, y = Gold)) + 
  geom_point() +
  facet_wrap(~NOC, scales = 'free') 

# Facet wrap can be added as a layer to any type of geom
```


## Contour/tile/image/level plots:

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
data(faithfuld)

#faithfuld is a data transofrmation of faithful, which is a 'built in' data set describing the time between eruptions and the duration of the eruption for the Old Faithful geyser in Yellowstone National Park, Wyoming, USA.
# - eruptions: eruption time in mins
# - waiting: waiting time to next eruption in mins
# - density: 2d density estimate

head(faithfuld)

ggplot(data = faithfuld) +
  aes(x = eruptions, y = waiting) +
  geom_raster(aes(fill = density)) 


ggplot(data = faithfuld) +
  aes(x = eruptions, y = waiting, fill = density) +
  geom_contour(aes(z = density)) 
```

## Heatmaps

```{r}
df <- expand.grid(x = 0:5, y = 0:5)
df$z <- sample(10:100, nrow(df), replace = TRUE) * sqrt(df$x) + df$y * 10

df %>%
  ggplot() +
  aes(x = x, y = y, fill = z) +
  geom_tile() +
  coord_equal()

```
# aes()

## Anatomy of `aes()`

So far, we've been mostly defining x and y variables in aes....

```{r, eval=FALSE}
# NOT run
ggplot(data = , aes(x = , y = , color = , linetype = , shape = , size = ))
```

These four aesthetic parameters (`color`, `linetype`, `shape`, `size`) can be used to show variation in *kind* (categories/factors) and variation in *degree* (numeric (ex: fill)).

I want to:

* Change the color of all the points in my x-y scatter plot to blue: supply 'color = ' OUTSIDE the `aes()` argument
* Change the color of all the points in my x-y scatter plot based on a variable in my data set: supply 'color = ' INSIDE the `aes()` argument

Parameters passed into `aes` should be *variables* in your dataset. (Can technically pass in new dataframes, but it gets messy and isn't recommended.)

Parameters passed to `geom_xxx` outside of `aes` should *not* be related to your dataset -- they apply to the whole figure.

```{r, fig.cap=" ", warning=FALSE}
usa_all_medals <- medal_counts %>%
  filter(NOC == "USA")

head(usa_all_medals)

ggplot(data = usa_all_medals, aes(x = Year, y = count, color = Medal, linetype = Medal)) +
            geom_line()
```
Long data is usually easier to work with in ggplot!

Note what happens when we specify the color parameter outside of the aesthetic operator. `ggplot2` views these specifications as invalid graphical parameters.

```{r, fig.cap=" ", warning=FALSE, error = TRUE}
ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
            geom_point(color = Medal)
ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
            geom_point(color = "Medal")
ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
            geom_point(color = "red")
```


# Changing options in ggplot2

`ggplot` handles options in additional layers.

### Labels

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + geom_point() +
  xlab(label = "Year") +
  ylab(label = "Number of Gold Medals Won") +
  ggtitle(label = "Cool Graph") # ggplot2
```

### Axis and point scales

Control options at a "global" level by not specifying within `aes`

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point() 

# Now with bigger points
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(size=3) 

# Now with tiny points
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(size=0.1) 
```

### Colors
```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(color = colors()[11]) # ggplot2
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(color = "#266F36") # ggplot2
```
The difference between 'color' and 'fill' can be confusing; 'color' refers to points and lines, while 'fill' refers to rectangles.

```{r}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_bar(stat = 'identity', 
                     color = "navyblue",
                     fill = 'lightpink') # ggplot2
```

### Point Styles and Widths

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(shape = 3) 
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(shape = "w") 
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_point(shape = "$", size=5) 
```

### Line Styles and Widths

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_line(linetype = 1) 
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_line(linetype = 'dashed')
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) +
            geom_line(linetype = 5, size = 2, color = 'blue') 
```

# Using aesthetics to highlight features

Differences in kind

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = northern_hem, aes(x = Year, y = Gold)) +
            geom_line(aes(linetype = NOC))


ggplot(data = northern_hem, aes(x = Year, y = Gold)) +
            geom_point(aes(shape = NOC, color = NOC))

# same as above but specifying the colors
ggplot(data = northern_hem, aes(x = Year, y = Gold)) +
  geom_point(aes(shape = NOC, color = NOC)) +
  scale_color_manual(values = c('darkblue', 'orangered', 'yellow4', 'pink3'))


# same as above but specifying the colors and the shape
ggplot(data = northern_hem, aes(x = Year, y = Gold)) +
  geom_point(aes(shape = NOC, color = NOC)) +
  scale_color_manual(values = c('darkblue', 'orangered', 'yellow4', 'pink3')) +
  scale_shape_manual(values = c(5, 6, 7, 8))


  
```
See named colors here: http://sape.inf.usi.ch/sites/default/files/ggplot2-colour-names.png
You can also supply color hexes. 

Differences in degree

```{r, fig.cap=" ", warning=FALSE, fig.show = "hold"}
ggplot(data = northern_hem, aes(x = Year, y = Silver)) +
            geom_point(aes(color = Gold))
ggplot(data = northern_hem, aes(x = Year, y = Silver)) +
            geom_point(aes(size = Gold))
```

Multiple non-coordinate aesthetics (differences in kind using color, degree using point size)

```{r, fig.cap=" ", warning=FALSE}
ggplot(data = northern_hem, aes(x = Year, y = Silver)) +
            geom_point(aes(size = Gold, color = NOC))
```



# Fitted lines and curves with `ggplot2`

```{r, fig.cap=" ", warning=FALSE, message=FALSE}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point()
```


```{r fig.cap=" ", warning=FALSE, message=FALSE}
# Add linear model (lm) smoother
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point() +
  geom_smooth(method = "lm", color = "orange") #by default, adds standard error

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# Add local linear model (loess) smoother, span of 0.75 (more smoothed)
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + geom_point() +
  geom_smooth(method = "loess", span = .75)

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# Add local linear model (loess) smoother, span of 0.25 (less smoothed)
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + geom_point() +
  geom_smooth(method = "loess", span = .25)

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# Add linear model (lm) smoother, no standard error shading
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + geom_point() +
  geom_smooth(method = "lm", se = FALSE)

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# Add local linear model (loess) smoother, no standard error shading
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + geom_point() +
  geom_smooth(method = "loess", se = FALSE)

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# Add a local linear (loess) smoother for each medal, no standard error shading
ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "loess", se = FALSE)


```

# Manually defining lines

Sometimes you might want to add the results of a regression you did to a graph, or indicate particular values on your graph. The family of functions `geom_abline`, `geom_hline` and `geom_vline` are useful for this.

```{r}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point() +
  geom_smooth(method = "loess", se = FALSE) +
  geom_hline(yintercept = 100, color = "purple", linetype = "dashed") +
  annotate(geom = "text", label = "target medal count", 
           x = 1893, y = 100, vjust = -1, hjust = 0, color = "purple")
```

```{r}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point() +
  geom_smooth(method = "loess", se = FALSE) +
  geom_vline(xintercept = 1950, color = "purple", linetype = "dashed") +
  annotate(geom = "text", label = "start of new\nolympic athlete program", 
           x = 1950, y = 140, vjust = -0.4, hjust = 0.5, color = "purple", angle = 90)
```

```{r}
ggplot(data = usa_gold_medals, aes(x = Year, y = count)) + 
  geom_point() +
  geom_smooth(method = "loess", se = FALSE) +
  geom_abline(slope = 2.1, intercept = -4000, color = "purple") +
  annotate(geom = "text", label = "My Comptitor's Model", 
           x = 1960, y = 160, vjust = -0.4, hjust = 0.5, color = "purple") +
  annotate(geom = "text", label = "My Model", 
           x = 1983, y = 120, vjust = -0.4, hjust = 0.5, color = "blue") 
```

# Themes

So far our plots are perhaps prettier than Base R plots, but they're still kinda ugly...

Customizing the way the non-data component of your graph will usually require manipulating the `theme()` layer.

The (now familiar) grey background with white lines is part of the default ggplot theme, `theme_grey()`. ggplot2 comes with several built-in themes that can be easily added to a plot to change its appearance.

```{r default_themes}
# grey
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  theme_grey()

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# minimal
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  theme_minimal()

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# black and white
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  theme_bw()

```

Academic figures are often very minimalist... to fit that scheme try:

```{r fig.cap=" ", warning=FALSE, message=FALSE}

# classic
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  theme_classic() +
  #For this theme in particular, I like to resposition the axis to zero:
  scale_y_continuous(expand = c(0, 0, 0.1, 0.1))

```


```{r fig.cap=" ", warning=FALSE, message=FALSE}

# dark
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) +
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  theme_dark()

```

However, if you don't like these default themes, you can build your own. 

```{r}
usa_all_medals %>% 
  ggplot(aes(x = Year, y = count)) +
  geom_point(aes(color = Medal)) + 
  scale_y_continuous(expand = c(0, 0, 0.1, 0.1))+
  geom_smooth(aes(color = Medal), method = "lm", se = FALSE) +
  ylab("Number of medals") +
  theme(legend.position = 'top',
        # most theme elements are specified as a line element, a rect element,
        # a text element, or a blank element to remove it entirely
        legend.background = element_blank(),
        legend.key = element_blank(),
        
        plot.background = element_rect(fill = '#fcfbed'),
        panel.background = element_rect(fill = '#fcfbed'),
        
        axis.line = element_line(color = '#633c01'),
        
        axis.text = element_text(size = 14),
        legend.text = element_text(size = 12),
        legend.title = element_text(size = 14), 
        axis.title = element_text(size = 16),
        
        panel.grid = element_line(color = '#8a7453'),
        panel.grid.major = element_line(linetype = 'dashed'),
        panel.grid.minor = element_blank()) +
  scale_color_manual(values = c('orangered', 'darkturquoise', 'navyblue'))

```
You can also download ggplot theme libraries. For example, the package `ggthemes` has everything you need to copycat the Economist graphing style: https://www.ggplot2-exts.org/ggthemes.html 

# The ordering of layers matters!

```{r layeroverride}

# Compare boxplot first, then jitter
iris %>%
  ggplot() +
  aes(x = Species, y = Sepal.Length) +
  geom_boxplot(outlier.shape = NA) +
  geom_jitter(width = 0.3)                  # jitter: point with position dodging

# versus jitter first, then boxplot
iris %>%
  ggplot() +
  aes(x = Species, y = Sepal.Length) +
  geom_jitter(width = 0.3) +
  geom_boxplot(outlier.shape = NA)


```

# Combining Multiple Plots

There are many packages you can explore to 

## GridExtra

* `ggplot2` graphs can be combined using the *`grid.arrange()`* function in the **`gridExtra`** package

```{r, warning=FALSE, fig.cap=" ", fig.width = 8, fig.height = 5}
# Initialize gridExtra library
library(gridExtra)

# Create 3 plots to combine in a table
plot1 <- ggplot(data = medal_counts_wide, aes(x = Year, y = Gold)) +
  geom_point(color = "gold") +
  geom_point(aes(x = Year, y = Silver), color = "gray") +
  geom_point(aes(x = Year, y = Bronze), color = "brown") +
  ylab("Medals")
plot2 <- ggplot(data = usa_all_medals, aes(x = Year, y = count)) +
      geom_line(aes(color = Medal)) + 
  theme(legend.position = 'left')
plot3 <- ggplot(data = northern_hem, aes(x = Year, y = Gold)) +
      geom_line(aes(linetype = NOC)) + 
  theme(legend.position = 'right')


# Call grid.arrange
grid.arrange(plot1, plot2, plot3, nrow=3, ncol = 1)
grid.arrange(plot1 + labs(tags = "A"), 
             plot2 + theme(legend.position = 'none') + labs(tags = "B"), 
             plot3 + theme(legend.position = 'none') + labs(tags = "C"), 
             layout_matrix = rbind(c(1,1), c(2, 3)))
```

## `patchwork`

* The `patchwork` package may be used to combine multiple `ggplot2` plots using
  a small set of operators similar to the pipe.
* This requires less syntax than using `gridExtra` and allows complex
  arrangements to be built more effortlessly.

```{r, warning=FALSE, fig.cap=" ", fig.width=12}
# Install and initialize patchwork library
# you will need `devtools` to install this one, if you don't have it yet: 
#install.packages('devtools')
#devtools::install_github("thomasp85/patchwork")
library(patchwork)

# use the patchwork operators
# stack plots horizontally
plot1 + plot2 + plot3

# stack plots vertically
plot1 / plot2 / plot3

# side-by-side plots with third plot below
(plot1 | plot2) / plot3

# side-by-side plots with a space in between, and a third plot below
(plot1 | plot_spacer() | plot2) / plot3

# stack plots vertically and alter with a single "gg_theme"
(plot1 / plot2 / plot3) & theme_classic()
```

Feel free to explore more at [https://github.com/thomasp85/patchwork](https://github.com/thomasp85/patchwork).

# Exporting

Two basic image types:

### **Raster/Bitmap** (.png, .jpeg)

Every pixel of a plot contains its own separate coding; not so great if you want to resize the image

```{r, eval=FALSE}
jpeg(filename = "example.jpg", width=, height=)
plot(x,y)
dev.off()
```

### **Vector** (.pdf, .ps)

Every element of a plot is encoded with a function that gives its coding conditional on several factors; great for resizing

```{r, eval=FALSE}
# NOT run
pdf(file = "example.pdf", width=, height=)
plot(x,y)
dev.off()
```

### Exporting with `ggplot`

```{r, eval=FALSE}
# NOT run

# Assume we saved our plot is an object called example.plot

ggsave(filename = "example.pdf", plot = example.plot, scale = , width = ,
       height = )
```

### Exporting via plots panel

Great for "quick" graphs, easy sharing. Not great for automatically generating graphs, reproducibly generating a graph.

# Inspiration

Some data is always presented the same way across studies (ex: DV vs IPRED will always include a scatter plot...). But sometimes you might find yourself unsure which `geom` will best communicate your data. The GG Gallery breaks down different graph types based on the types of data they communicate, and comes with sample code: https://www.r-graph-gallery.com/



# Extra Resources

`dplyr` and `tidyr` have many more functions to help you wrangle and manipulate your data. See the  [Data Wrangling Cheat Sheet](https://github.com/rstudio/cheatsheets/blob/master/data-transformation.pdf) for more.

These questions ask you to work with the gapminder dataset.

### Basics

1) Plot a histogram of life expectancy. 

```{r}
gap %>%
  ggplot() +
  aes(x = lifeExp) +
  geom_histogram()

```

2) Plot the life expectancy against gdpPercap. 

```{r}
gap %>%
  ggplot() +
  aes(x = gdpPercap) +
  geom_histogram() +
  facet_wrap(~continent) +
  scale_y_continuous(limits = c(0, 1000))

grid.arrange()

```

3) Clean up your scatterplot with a title and axis labels. Output it as a PDF and see if you'd be comfortable with including it in a report/paper.

```{r}
le_plot <- gap %>%
  ggplot() +
  aes(x = lifeExp) +
  geom_histogram() +
  xlab("Life Expectancy (years)") +
  ylab("Number of Countries")


ggsave("le_plot.pdf", le_plot)

```

4) Make a boxplot of life expectancy conditional on year.

```{r}

gap %>%
  ggplot() + 
  aes(x = factor(year), y = lifeExp) +
  geom_boxplot()


```

### Using the ideas

5) Create a trellis plot of life expectancy by gdpPercap scatterplots, one subplot per year. Use a 3x4 layout of panels in the plot.

```{r}


```

6) Plot life expectancy versus gdpPercap. Now plot so that different continents are in different colors. Use `scale_x_continuous()` to set the x-axis limits to be in the range from 100 to 50000.

```{r}
ggplot(gap, aes(x = gdpPercap, y = lifeExp)) +
  geom_point() +
  scale_x_continuous(limits = c(100, 50000)) +
  scale_y_continuous(limits = c(0, 100))

```

7) Figure out how to use the log-scale for gdpPercap, without manually calculating the log values.

```{r}
ggplot(gap, aes(x = gdpPercap, y = lifeExp)) +
  geom_point() +
  scale_x_log10(limits = c(100, 50000)) +
  scale_y_continuous(limits = c(0, 100)) +
  theme(axis.text.x = element_text(size = 20))

```

### Advanced

8) Create a "trellis" plot where, for a given year, each panel uses a) hollow circles to plot lifeExp as a function of log(gdpPercap), and b) a red loess smoother without standard errors to plot the trend. Turn off the grey background. Figure out how to use partially-transparent points to reduce the effect of the overplotting of points.

```{r}


```

9) Building on your graph in Question 8, change at least two different properties in the `theme()` layer to customize the graph to your liking.

```{r}

```