RMarkdown_Exercises_Set2_SOLUTIONS.Rmd

---
title: "Interactive R Markdown"
author: "Dustin Pluta"
date: "May 12, 2017"
output: 
  html_document:
    toc: true
    toc_float:
      collapsed: false
      smooth_scroll: true
    toc_depth: 3
    theme: spacelab
runtime: shiny
---


# Interactive Basics

1. [Here is a link to the RStudio guide to interactive R Markdown documents.](http://rmarkdown.rstudio.com/authoring_shiny.html?version=1.0.136&mode=desktop#overview)

2. By leveraging the power of [Shiny](http://shiny.rstudio.com/), we can make R Markdown documents interactive!

  * [Shiny Widget Gallery](http://shiny.rstudio.com/gallery/widget-gallery.html)
  * [Shiny App Gallery](https://shiny.rstudio.com/gallery/)

***
***

<br><br>

## Iris Examples

```{r, echo = FALSE, message = FALSE}
library(dplyr)
library(ggplot2)
numericInput("rows", "How many Flowers?", 5)
textInput("species", "Which Species?", "virginica")

renderTable({
  head(filter(iris, Species == input$species), input$rows)
})
```

<br><br>

```{r, echo = FALSE}
library(ggplot2)
library(png)
radioButtons("radio", label = h4("Choose a Plot"),
    choices = list("Violin Plot" = 1, "Scatterplot" = 2, "Iris" = 3), selected = 1)

renderPlot(
  if (input$radio == 1) {
    ggplot(data=iris, aes(x=Species, y=Petal.Width)) +
      geom_violin(aes(fill=Species)) + geom_point()
  } else if (input$radio == 2) {
    ggplot(iris, aes(Petal.Length, Petal.Width, color =        Species)) + geom_point()
  } else if (input$radio == 3) {
    rasterImage(readPNG('Figures/iris.png'), xleft = 0.1, xright = 1, ytop = 1, ybottom = 0)
  }
)
```

***
***

<br><br>

## Exercises with the IMDB Data

Let's analyze the IMDB data interactively.

First, read the data as before:

```{r read_imdb, results = "hide"}
imdb <- read.csv("Data/movie_metadata.csv", stringsAsFactors = FALSE)
colnames(imdb)
```

***

<br><br>

### Problem 1

1. Draw a histogram of the log movie budgets and include a slider to let you select the number of bins.  The example [here](http://rmarkdown.rstudio.com/authoring_shiny.html?version=1.0.136&mode=desktop#inputs_and_outputs) shows an example.  Use the base `hist()` for this part instead of `ggplot`.

```{r, echo = FALSE}
library(magrittr)
sliderInput("bins", "Number of bins:", min = 1, max = 50, value = 30)

renderPlot({
  bins <- seq(2, 10, length.out = input$bins + 1)
  x <- filter(imdb, !is.na(budget)) %>%
    transmute(log_budget = log(budget, 10)) %>%
    select(log_budget) %$%
    hist(log_budget, breaks = input$bins, col = 'darkgray', border = 'white')
})
```

***

### Problem 2

2. Revisit the plot of Nicolas Cage movies by year and IMDB score.  Generate this plot and include `numericInput` boxes to select the year range to be plotted.  Bonus points if you rewrite using `dplyr` functions.

Here is the code from the `ggplot` section:

```{r nic_cage, echo = TRUE, eval = FALSE}
cage <- imdb[imdb$actor_1_name == 'Nicolas Cage', ]
cage$actor_1_name <- as.character(cage$actor_1_name)
ggplot(data=cage, aes(x=title_year, y=imdb_score, label=movie_title)) +
  geom_point(alpha=0.5) + 
  geom_text(fontface='italic', size=2, vjust=1, nudge_y=0.1) + 
  geom_smooth()
```

__Solution__

```{r cage, echo = FALSE}
numericInput("min_year", h4("Min Year"), 1984)
numericInput("max_year", h4("Max Year"), 2017)

renderPlot({
cage <- filter(imdb, actor_1_name == 'Nicolas Cage') %>% 
  filter(title_year > input$min_year & title_year < input$max_year)
cage$actor_1_name <- as.character(cage$actor_1_name)
ggplot(data=cage, aes(x=title_year, y=imdb_score, label=movie_title)) +
  geom_point(alpha=0.5) + 
  geom_text(fontface='italic', size=2, vjust=1, nudge_y=0.1) + geom_smooth()
})
```

***

<br><br>

### Problem 3

3. Add a `textInput` to allow the choice of actor to plot for the previous plot.

__Solution__

```{r P3, echo = FALSE}

numericInput("P3_min_year", h4("Min Year"), 1984)
numericInput("P3_max_year", h4("Max Year"), 2017)
textInput("actor", h4("Choose Actor"), "Nicolas Cage")

renderPlot({
cage <- filter(imdb, actor_1_name == input$actor) %>% 
  filter(title_year > input$P3_min_year & title_year < input$P3_max_year)
cage$actor_1_name <- as.character(cage$actor_1_name)
ggplot(data=cage, aes(x=title_year, y=imdb_score, label=movie_title)) +
  geom_point(alpha=0.5) + 
  geom_text(fontface='italic', size=2, vjust=1, nudge_y=0.1) + geom_smooth()
})
```

***

<br><br>

### Problem 4

4. Modify the following code to allow a movie name to be entered for highlighting using a `textInput` box.

```{r message = FALSE, warning = FALSE}
ggplot(data=imdb, aes(x=log10(budget), y=imdb_score)) +
geom_point(aes(colour=gross), alpha=0.5) +
geom_smooth() +
scale_color_continuous(name='Box office gross', breaks = c(2e+8, 4e+8, 6e+8),
labels = c('2 million', '4 million', '6 million'),
low = 'blue', high = 'red') +
annotate('point', x=8.3, y=8.3, colour='red', size=4) +
annotate('text', x=8.3, y=8.6, label='Toy Story 3', fontface='italic', size=6, angle=30) +
theme_bw() +
labs(title='IMDB Movies') +
theme(plot.title=element_text(size=rel(2), colour='blue')) +
theme(legend.position=c(0.9, 0.2))
```

```{r, message = FALSE, warning = FALSE}

textInput("movie_title", h4("Enter Movie Title"), "Toy Story 3")

renderPlot({
  movie_index <- grep(input$movie_title, imdb$movie_title)
  selected_x <- log(imdb$budget[movie_index], 10)
  selected_y <- imdb$imdb_score[movie_index]
  
  ggplot(data=imdb, aes(x=log10(budget), y=imdb_score)) +
  geom_point(aes(colour=gross), alpha=0.5) +
  geom_smooth() +
  scale_color_continuous(name="Box Office Gross", breaks = c(2e+8, 4e+8, 6e+8),
  labels = c('2 million', '4 million', '6 million'),
  low = 'blue', high = 'red') +
  annotate('point', x=selected_x, selected_y, colour='red', size=6) +
  annotate('text', x=selected_x, y=selected_y, label=imdb$movie_title[movie_index], fontface='bold', size=10) +
  theme_bw() +
  labs(title='IMDB Movies') +
  theme(plot.title=element_text(size=rel(2), colour='blue')) +
  theme(legend.position=c(0.9, 0.2))
})
```


### Using `plotly`

- Install the package `plotly`.  We can include plotly plots using

```{r}
library(plotly)
dsamp <- diamonds[sample(nrow(diamonds), 1000), ]

renderPlotly({
  p <- qplot(carat, price, data=dsamp, colour=clarity)
  ggplotly(p)  
})
```

### Problem 5

5. Redo your previous IMDB data plots to use `plotly`.

    1. Change `renderPlot` to `renderPlotly`.
    
    2. Assign the `ggplot2` graph object to a variable, call it `p`.
    
    3. Add the line `ggplotly(p)` at the end of the `renderPlotly` block.

```{r, message = FALSE, warning = FALSE}
library(plotly)
textInput("movie_title", h4("Enter Movie Title"), "Toy Story 3")

renderPlotly({
  movie_index <- grep(input$movie_title, imdb$movie_title)
  selected_x <- log(imdb$budget[movie_index], 10)
  selected_y <- imdb$imdb_score[movie_index]
  
  p <- ggplot(data=imdb, aes(x=log10(budget), y=imdb_score)) +
  geom_point(aes(colour=gross), alpha=0.5) +
  geom_smooth() +
  scale_color_continuous(name="Box Office Gross", breaks = c(2e+8, 4e+8, 6e+8),
  labels = c('2 million', '4 million', '6 million'),
  low = 'blue', high = 'red') +
  annotate('point', x=selected_x, selected_y, colour='red', size=6) +
  annotate('text', x=selected_x, y=selected_y, label=imdb$movie_title[movie_index], fontface='bold', size=10) +
  theme_bw() +
  labs(title='IMDB Movies') +
  theme(plot.title=element_text(size=rel(2), colour='blue')) +
  theme(legend.position=c(0.9, 0.2))
  
  ggplotly(p)
})
```