Please remote connect to your desktop computer

1. Open the Start menu  *Click the Window’s logo on the bottom left of the screen
2. Select Remote Desktop Connection
3. Enter w7-your7digit#
4. Press Connect

Open your RStudio project

1. Open your project folder from last week.
2. Double click the .Rproj file to open RStudio.
3. Ready yourself for an adventure.

CAMP schedule

Day 2 review

Reminder: Add your packages to the top of each new script.

library(readr)
library(dplyr)
library(ggplot2)

# Your code starts here.

Let’s load the IMDB movie data again.

Surprise! The data is now in an Excel file.

Line charts

Now let’s look at the total gross of movies over time in millions of dollars.

gross_by_year <- movies %>% 
                         group_by(title_year) %>% 
                         summarize(total_gross = sum(gross_mil, na.rm = T)) %>% 
                         ungroup()

ggplot(gross_by_year, aes(x = title_year, y = total_gross)) + 
  geom_line() 

gross_by_year <- movies %>% 
                   group_by(title_year, content_rating) %>% 
                   summarize(total_gross = sum(gross_mil, na.rm = T)) %>% 
                   ungroup()

ggplot(gross_by_year, aes(x = title_year, y = total_gross)) + 
  geom_line(aes(color = content_rating)) 

Title and axis labels

ggplot(gross_by_year, aes(x = title_year, y = total_gross, color = content_rating)) + 
  geom_line() +
  labs(title = "Is the movie industry making more money now?",
       x     = "Year",
       y     = "Gross in millions of U.S. $$$$")

ggplot(gross_by_year, aes(x = title_year, y = total_gross, color = content_rating)) + 
  geom_line() +
  labs(title = "The movie industry is making more money over time",
       subtitle = "What happened in 2012? Netflix? YouTube? New books?",
       x        = "Year",
       y        = "Gross in millions of U.S. $$$$",
       caption  = "Awesome chart made by Derek. December 2017.")

Let’s compare how much money movies made by content rating. Use y = gross_mil for the y-axis.

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  geom_boxplot()

Transform to Log scale.

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  geom_boxplot() +
  scale_y_log10()

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  geom_boxplot(outlier.color = "tomato") + 
  scale_y_log10()

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  geom_boxplot(outlier.color = "tomato") + 
  scale_y_log10() + 
  geom_point(alpha = 0.1)

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  geom_boxplot(outlier.color = NA) + 
  scale_y_log10() + 
  geom_jitter(alpha = 0.1)

Rearrange the layer order to put the boxplots on top.

ggplot(movies, aes(x = content_rating, y = gross_mil)) +
  scale_y_log10() + 
  geom_jitter(alpha = 0.1) + 
  geom_boxplot(outlier.color = NA, alpha = 0.8)

Exercise

Make boxplots comparing the range of gross_mil between movie_color (black and white films vs. color films).

Show solution

ggplot(movies, aes(x = movie_color, y = gross_mil)) +
  scale_y_log10() + 
  geom_jitter(alpha = 0.1) + 
  geom_boxplot(outlier.color = NA, alpha = 0.8)

Johnny Depp movies

We received questions about creating a new column with mutate() based on the values of an existing column.

To learn how to do this, let’s make a boxplot of how much movies make that starred Johnny Depp vs. those that didn’t (big mistake!). First we need to learn about ifelse().

ifelse() makes a comparison and then returns a value depending on the result of the comparison. The comparison below looks to see whether a movie’s #1 actor was Johnny Depp, if it was then ifelse() returns Yes, if not it returns No.

movies <- movies %>%
          mutate(stars_johnny_depp = ifelse(actor_1_name == "Johnny Depp", "Yes", "No"))

ggplot(movies, aes(x = stars_johnny_depp, y = gross_mil)) +
  geom_boxplot() +
  scale_y_log10()

Use a bar chart to compare imdb_score by content_rating

imdb_score_by_content <- movies %>% 
                           group_by(content_rating) %>% 
                           summarize(mean_score = mean(imdb_score, na.rm = T))

ggplot(imdb_score_by_content, aes(x = content_rating, y = mean_score)) + geom_col() 

Note: By default geom_bar() simply counts the number of values in each category. If you want to show the height of each bar at a specific value in the data, use geom_bar(stat = "identity"). The function geom_col() is equivalent to geom_bar(stat = "identity").

Labels for each bar

There’s a geom_label we can use to label each column.

ggplot(imdb_score_by_content, aes(x = content_rating, y = mean_score)) +
  geom_col() +
  geom_label(aes(label = mean_score))

Hint: Update to label = round(mean_score, digits = 1)) to make labels legible.

budget_by_content <- movies %>% 
                     group_by(content_rating) %>% 
                     summarize(mean_budget         = mean(budget, na.rm = T) / 1000000,
                               mean_budget_round   = round(mean_budget, digits = 1)) 

ggplot(budget_by_content, aes(x = content_rating, y = mean_budget_round)) +
  geom_bar(stat = "identity") +
  geom_label(aes(label = mean_budget_round))

Just for fun…

movies <- mutate(movies, 
                 profit     = gross - budget,
                 profit_mil = profit / 1000000)

ggplot(movies, aes(x = title_year,  y = profit)) + geom_jitter(alpha = 0.5)

ggplot(movies, aes(x = title_year,  y = profit)) +
  geom_jitter(alpha = 0.5) + 
  geom_label(aes(x = title_year,  
                 y = profit,
                 label = ifelse(profit > quantile(movies$profit, 0.997, na.rm = T) |
                                profit < quantile(movies$profit, 0.003, na.rm = T),
                                movie_title,
                                NA)),
             alpha = 0.85, 
             nudge_y = 3) +
  labs(title  = "Which movies made and lost the most $$$ ?",
       y      = "Profit in U.S. dollars",
       x      = "year")

library(hrbrthemes)  #install.packages(hrbrthemes)
library(scales)      # For the $ dollar_format() labels

ggplot(movies, aes(x = title_year,  y = profit)) + 
  geom_label(aes(label = movie_title, 
                 fill  = profit > 0),
                 show.legend = F,
                 alpha = 0.9) + 
  scale_y_continuous(labels = dollar_format()) + 
  labs(title    = "Did your favorite film make or lose $$$ ?",
       subtitle = "Profit of top 5,000 rated films on IMDB",
       y        = "Profit in U.S. dollars",
       x        = "year",
       caption  = "Brought to you by the #R-Cats") +
  theme_ipsum_rc() + 
  theme(plot.title = element_text(size = 24)) # Increase title size

How many movies were made in each content rating?

Hint: use + coord_flip()

Are the number of movies increasing for G, PG, PG-13 and R rated films?

Hint: Use %in% to match on multiple values when you filter().

Since 2012, have the total Facebook likes for a movie been associated with the number of cast Facebook likes?

Hint: We made a similar chart in Day 2.

What is the trend in the total gross of G and PG rated films?

Hint: Create a summarized table grouped by title_year and content_rating

Since 2012, which content rating has had the highest median number of Facebook likes?

Which type of movie has the shortest running time: G, PG-13, or R?

Movie budgets of longer movies.

Hint: Create a new column called longer_than_2hrs and set it equal to duration > 120

Hint 2: Use scale_y_log10()

They say attention spans are dwindling. Have movie durations been decreasing since 1965?

Hint: Use geom_jitter()

Since 1990, what was the average IMDB score for each content rating?

Hint: Summarize grouped by content_rating and title_year.

Hint 2: Use geom_tile(aes(fill = mean_score)).

filter(movies, title_year >= 1990) %>% 
  group_by(content_rating, title_year) %>% 
  summarize(mean_score = mean(imdb_score, na.rm = T)) %>% 
  ggplot(aes(y = content_rating, x = title_year)) +
  geom_tile(aes(fill = mean_score), color = "white") +   
  #theme_minimal() +
  labs(title = "Average IMDB score by content rating")

Show the distribution of gross for each content rating.

#install.packages("ggridges")
library(ggridges)

ggplot(movies, aes(x = gross_mil, y = content_rating, fill = content_rating)) + # x/y reveresed from boxplots
  geom_density_ridges() +
  scale_x_log10() + 
  scale_fill_brewer() + 
  labs(title = "Which rating makes more $$$ ?",
       x = "Gross in millions of dollars",
       y = "Content rating")

Frequent plotting questions

• How to modify the gridlines behind your chart?
  • Try the different themes at the end of this lesson: theme_light() or theme_bw().
  • Or modify the color and size with theme(panel.grid.minor = element_line(colour = "white", size = 0.5)).
  • There’s even theme_excel()
• How do you set the x and y scale manually?
  • Here is an example with a scatter plot: ggplot() + geom_point() + xlim(beginning, end) + ylim(beginning, end)
  • Warning: Values above or below the limits you set will not be shown. This is another great way to lie with data.
• How do you get rid of the legend if you don’t need it?
  • geom_point(aes(color = facility_name), show.legend = FALSE)
  • The R cookbook shows a number of ways to get rid of legends: http://www.cookbook-r.com/Graphs/Legends_(ggplot2)/
• I only like dashed lines. How do you change the linetype to a dashed line?
  • geom_line(aes(color = facility_name), linetype = "dashed")
  • You should also try "dotted" and "dotdash", or maybe"twodash" if you really want to go wild. -How many colors are there in R? How did you know hotpink was a color?
  • There is an R color cheatsheet: https://www.nceas.ucsb.edu/~frazier/RSpatialGuides/colorPaletteCheatsheet.pdf
  • And some basic color names here: http://www.r-graph-gallery.com/42-colors-names/
  • This web tool will give you palette ideas and color Hex numbers: http://colorbrewer2.org/#type=sequential&scheme=BuGn&n=3
  • color palettes() There are color palettes for your favorite movies, just fun palettes that you might like library(viridis), color-blind palettes, and topography palettes.
• What are some handy keyboard shortcuts in R? How do I find them?
  • There is a shortcut cheat sheet online here: https://www.rstudio.com/wp-content/uploads/2016/01/rstudio-IDE-cheatsheet.pdf
  • Or go to Help > Keyboard Shortcuts Help

Good luck!