--- title: "STAT 331 Week 4 Day 2 Handout" format: html embed-resources: true --- ```{r} #| label: setup #| message: false #| echo: false library(tidyverse) library(readxl) library(janitor) library(ggridges) ``` ## Data ::: callout-warning For this handout, you need to download TS_data.xlsx from the Week 4 Schedule Canvas page and save it on your computer. ::: Change the file path below as needed to read in the imdb_data.Rdata file that we used on Monday. ```{r} load(file = "data/imdb_data.Rdata") ``` Save the `TS_data.xlsx` file to a reasonable directory, and change the file path below as needed to read it in. ```{r} eras_data <- read_excel(path = "data/TS_data.xlsx") ``` Change the file path below to read in the military spending data that we used for PA3. ```{r} military <- read_xlsx("data/SIPRI-Milex-data-1949-2024.xlsx", sheet = "Share of Govt. spending", skip = 7) ``` ## Variable Names in R a. How do column names need to be referenced in `dplyr` if they start with a number or include spaces? b. What does the `clean_names()` function from the `janitor` package do? ```{r} military |> clean_names() |> colnames() ``` ## Advanced Joins a. What were the active years of each director? ```{r} ``` b. What happens for directors that have multiple movies in the join below? ```{r} directors |> inner_join(movies_directors, join_by(id == director_id)) ``` c. When do you need to join on multiple keys? d. What is the general syntax for joining on multiple keys? ## Working with Categorical Variables (Factors) a. What are the two general purposes for factor variables (in R)? b. What function have you used previously to create a factor variable? ### `fct()` a. The `fct()` function in **forcats** serves the same purpose as what function in base R? b. How are the levels automatically ordered with the `fct()` function? How is this different from base R? ```{r} eras_data |> mutate(Album = fct(Album)) |> pull(Album) ``` c. Edit the code above to just output the levels of the `Album` variable. d. What does including the `levels = ` argument do in `fct()`? ```{r} eras_data |> mutate(Album = fct(Album, levels = c("Fearless","Speak Now","Red", "1989", "Reputation","Lover", "Folklore", "Evermore","Midnights"))) |> pull(Album) ``` e. What happens if you forget to provide an album to `levels = ` that exists in the data in the code above? ### `fct_recode()` a. What does this function do? When would you want to use it? b. What happens to non-specified levels in `fct_recode()`? c. Change the level "Red" to "Red!" and "Fearless" to "Fearless - Taylors Version" in the `Album` variable: ```{r} ``` ### `fct_collapse()` a. What does this function do? When would you want to use it? ```{r} eras_data |> mutate(Genre = fct_collapse(.f= Album, "country pop" = c("Fearless"), "pop rock" = c("Speak Now","Red"), "electropop" = c("1989","Reputation","Lover"), "folk pop" = c("Folklore","Evermore"), "alt-pop" = "Midnights")) ``` b. Why would you want to use `fct_collapse()` rather than `case_when()`? c. What happens to levels of the variable `.f = ` that you do not include in `fct_collapse()`? ```{r} eras_data |> mutate(genre = fct_collapse(.f= Album, "country pop" = c( "Fearless"), "pop rock" = c("Speak Now","Red"), "electropop" = c("1989","Reputation","Lover"), "folk pop" = c("Folklore","Evermore") )) |> pull(genre) |> levels() ``` d. Edit the code below to fix this, so that the Midnights album is categorized as "other" in the `genre` variable. ```{r} eras_data |> mutate(genre = fct_collapse(.f= Album, "country pop" = c( "Fearless"), "pop rock" = c("Speak Now","Red"), "electropop" = c("1989","Reputation","Lover"), "folk pop" = c("Folklore","Evermore") )) |> pull(genre) |> levels() ``` ### `fct_relevel()` a. What does this function do? When would you want to use it? ```{r} eras_data |> mutate(Album = fct_relevel(.f = Album, c("Fearless","1989","Taylor Swift", "Speak Now","Red","Midnights","Reputation"))) |> pull(Album) |> levels() ``` b. What happens with levels that are not specified? ### `fct_infreq()` a. What does `fct_infreq()` do? When would you want to use it? ```{r} eras_data |> ggplot() + geom_bar(aes(y = fct_infreq(Album)), fill = "#A5C9A5") + theme_minimal() + labs(x = "Number of Songs", y = "", subtitle = "Album", title = "Songs Played on the Eras Tour") ``` ### `fct_reorder()` a. What does `fct_reorder()` do? In what setting would you use it? ```{r} # you may need to install this package! library(ggridges) eras_data |> ggplot(aes(x = Length, y = fct_reorder(.f = Album, .x = Length, .fun = mean), fill = Album)) + geom_density_ridges() + theme_minimal() + theme(legend.position = "none")+ labs(x = "Song Length (mins)", y = "", subtitle = "Album", title = "Songs Played on the Eras Tour") ``` ### `fct_reorder2()` a. What does `fct_reorder2()` do? In what setting would you use it? b. What are the arguments to `fct_reorder2()`? c. Let's improve the plot that we made way back in week 2 of the number of listings over time for five cities in Texas. Use `fct_reorder2()` to reorder the legend below so that the legend aligns with the order of the lines in the plot. ```{r} data(txhousing) txhousing |> filter(city %in% c("Dallas","Fort Worth", "Austin", "Houston", "El Paso"), !is.na(listings)) |> ggplot() + geom_line(mapping = aes(x = date, y = listings, color = city)) ```