Data Wrangling with {dplyr}

class: title-slide, center, middle

# Data Wrangling with `dplyr`

---
background-image: url(images/hex/tidyverse.png)
background-position: 90% 5%
background-size: 10%

# Tidyverse

There are a few key ideas to be aware of about how the tidyverse works in general before we dive into `dplyr`

1. Packages are designed to be like **grammars** for their task. You can string these grammatical elements together to form more complex statements, just like with language.

1. The first argument of (basically) every function is `data`. This is very handy, especially when it comes to piping.

1. Variable names are usually not quoted (read more [here](https://tidyselect.r-lib.org/reference/language.html))

---
background-image: url(images/hex/palmerpenguins.png)
background-position: 90% 5%
background-size: 10%

# Palmer penguins

```r
library(palmerpenguins)
glimpse(penguins)
```

```
## Rows: 344
## Columns: 8
## $ species           <fct> Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adelie, Adel…
## $ island            <fct> Torgersen, Torgersen, Torgersen, Torgersen, Torgersen, Torgersen, Torgersen,…
## $ bill_length_mm    <dbl> 39.1, 39.5, 40.3, NA, 36.7, 39.3, 38.9, 39.2, 34.1, 42.0, 37.8, 37.8, 41.1, …
## $ bill_depth_mm     <dbl> 18.7, 17.4, 18.0, NA, 19.3, 20.6, 17.8, 19.6, 18.1, 20.2, 17.1, 17.3, 17.6, …
## $ flipper_length_mm <int> 181, 186, 195, NA, 193, 190, 181, 195, 193, 190, 186, 180, 182, 191, 198, 18…
## $ body_mass_g       <int> 3750, 3800, 3250, NA, 3450, 3650, 3625, 4675, 3475, 4250, 3300, 3700, 3200, …
## $ sex               <fct> male, female, female, NA, female, male, female, male, NA, NA, NA, NA, female…
## $ year              <int> 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007, 2007…
```

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# dplyr

`dplyr` is a grammar of data manipulation, providing a consistent set of core verbs that help you solve the most common data manipulation challenges

--
***

**Manipulating observations**

+ `filter()` picks cases based on their values.

+ `arrange()` changes the ordering of the rows.

--
***

**Manipulating variables**

+ `select()` picks variables based on their names.

+ `mutate()` adds new variables that are functions of existing variables

--
***

**Summarizing data**

+ `summarise()` reduces multiple values down to a single summary.

---
# Concept Map: `dplyr`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
background-image: url(images/hex/magrittr.png)
background-position: 90% 5%
background-size: 10%

# A review of pipes

`x %>% f(y)` is equivalent to `f(x, y)`

--
***

R Code

```r
penguins %>% 
  filter(species == "Gentoo") %>% 
  select(bill_length_mm, bill_depth_mm) %>% 
  arrange(desc(bill_length_mm))
```

Translated into English

```r
start with penguins data *AND THEN*
  filter to include only observations from Gentoo penguins *AND THEN*
  select only the columns `bill_length_mm` and `bill_depth_mm` *AND THEN*
  arrange observations by descending order of `bill_length_mm`
```

--
***
Read more on piping: [https://magrittr.tidyverse.org/reference/pipe.html](https://magrittr.tidyverse.org/reference/pipe.html)

---
background-image: url(images/hex/magrittr.png)
background-position: 90% 5%
background-size: 10%

# Concept Map: `%>%`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
class: inverse, center, middle

# Manipulating observations <br> (rows)

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `filter()`

### Subset observations (rows) with `filter()`

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `filter()`

.footnote[Artwork by [@allison_horst](https://twitter.com/allison_horst)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `filter()`

### Logical operators

.footnote[Source: [R for Data Science](https://r4ds.had.co.nz/transform.html#filter-rows-with-filter)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `filter()`

### Comparisons

<table>
 <thead>
  <tr>
   <th style="text-align:left;"> Operator </th>
   <th style="text-align:left;"> Description </th>
   <th style="text-align:left;"> Usage </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> &lt; </td>
   <td style="text-align:left;"> less than </td>
   <td style="text-align:left;"> x &lt; y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> &lt;= </td>
   <td style="text-align:left;"> less than or equal to </td>
   <td style="text-align:left;"> x &lt;= y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> &gt; </td>
   <td style="text-align:left;"> greater than </td>
   <td style="text-align:left;"> x &gt; y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> &gt;= </td>
   <td style="text-align:left;"> greater than or equal to </td>
   <td style="text-align:left;"> x &gt;= y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> == </td>
   <td style="text-align:left;"> exactly equal to </td>
   <td style="text-align:left;"> x == y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> != </td>
   <td style="text-align:left;"> not equal to </td>
   <td style="text-align:left;"> x != y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> %in% </td>
   <td style="text-align:left;"> group membership </td>
   <td style="text-align:left;"> x %in% y </td>
  </tr>
  <tr>
   <td style="text-align:left;"> is.na </td>
   <td style="text-align:left;"> is missing </td>
   <td style="text-align:left;"> is.na(x) </td>
  </tr>
  <tr>
   <td style="text-align:left;"> !is.na </td>
   <td style="text-align:left;"> is not missing </td>
   <td style="text-align:left;"> !is.na(x) </td>
  </tr>
</tbody>
</table>

.footnote[Source: [Alison Hill](https://share-blogdown.netlify.app/slides/02-slides.html#15)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `filter()`

.panelset[

.panel[.panel-name[Arguments]

<code class ='r hljs remark-code'>filter(<span style="color:cornflowerblue">.data</span>, <span style="color:red">...</span>)</code>

**.blue[.data]** .blue[= a data frame or tibble]

**.red[. . .]**  .red[= Expressions that return a logical value, and are defined in terms of the variables in .blue[.data]. <br><br> If multiple expressions are included, they are combined with the & operator. Only rows for which all conditions evaluate to TRUE are kept. ]
]

.panel[.panel-name[Example 1]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55)
```

```
## # A tibble: 3 x 8
##   species island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g sex    year
##   <fct>   <fct>           <dbl>         <dbl>             <int>       <int> <fct> <int>
## 1 Gentoo  Biscoe           59.6            17               230        6050 male   2007
## 2 Gentoo  Biscoe           55.9            17               228        5600 male   2009
## 3 Gentoo  Biscoe           55.1            16               230        5850 male   2009
```

]

.panel[.panel-name[Example 2]

```r
penguins %>% 
  filter(species %in% c("Adelie", "Gentoo"),
         island %in% c("Dream", "Torgersen"))
```

```
## # A tibble: 108 x 8
##    species island    bill_length_mm bill_depth_mm flipper_length_mm body_mass_g sex     year
##    <fct>   <fct>              <dbl>         <dbl>             <int>       <int> <fct>  <int>
##  1 Adelie  Torgersen           39.1          18.7               181        3750 male    2007
##  2 Adelie  Torgersen           39.5          17.4               186        3800 female  2007
##  3 Adelie  Torgersen           40.3          18                 195        3250 female  2007
##  4 Adelie  Torgersen           NA            NA                  NA          NA <NA>    2007
##  5 Adelie  Torgersen           36.7          19.3               193        3450 female  2007
##  6 Adelie  Torgersen           39.3          20.6               190        3650 male    2007
##  7 Adelie  Torgersen           38.9          17.8               181        3625 female  2007
##  8 Adelie  Torgersen           39.2          19.6               195        4675 male    2007
##  9 Adelie  Torgersen           34.1          18.1               193        3475 <NA>    2007
## 10 Adelie  Torgersen           42            20.2               190        4250 <NA>    2007
## # … with 98 more rows
```
]
]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Concept Map: `filter()`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `arrange()`

### Arrange rows by column values with `arrange()`

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `arrange()`

.panelset[

.panel[.panel-name[Arguments]

<code class ='r hljs remark-code'>arrange(<span style="color:cornflowerblue">.data</span>, <span style="color:red">...</span>)</code>

**.blue[.data]** .blue[= a data frame or tibble]

**.red[. . .]**  .red[= Variables to sort by. Use `desc()` to sort a variable in descending order.]
]

.panel[.panel-name[Example 1]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g)
```

```
## # A tibble: 3 x 8
##   species island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g sex    year
##   <fct>   <fct>           <dbl>         <dbl>             <int>       <int> <fct> <int>
## 1 Gentoo  Biscoe           55.9            17               228        5600 male   2009
## 2 Gentoo  Biscoe           55.1            16               230        5850 male   2009
## 3 Gentoo  Biscoe           59.6            17               230        6050 male   2007
```

]

.panel[.panel-name[Example 2]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(desc(body_mass_g))
```

```
## # A tibble: 3 x 8
##   species island bill_length_mm bill_depth_mm flipper_length_mm body_mass_g sex    year
##   <fct>   <fct>           <dbl>         <dbl>             <int>       <int> <fct> <int>
## 1 Gentoo  Biscoe           59.6            17               230        6050 male   2007
## 2 Gentoo  Biscoe           55.1            16               230        5850 male   2009
## 3 Gentoo  Biscoe           55.9            17               228        5600 male   2009
```
]
]

---
class: inverse

# Your turn 1

1. Import the file `pragmatic_scales_data.csv` (use best practices of a project-oriented workflow). Save it to an object called `ps_data`. This time we'll save it as a `tibble` (which I've done for you).

1. Filter rows for cases in the "No Label" `condition` and arrange the resulting observations by descending order of `age`.

1. Select observations from the “No Label” condition for kids 3 years old or younger.

1. Select observations for kids between the ages of 3 and 4. Save the result to a new object called `ps_filtered`. (For an extra challenge, look up the documentation for `between()` and use this function instead of comparison operators).

---
class: solution

# Solution

.panelset[
.panel[.panel-name[Q1]

```r
# Q1.
ps_data <- rio::import(here::here("data", "pragmatic_scales_data.csv")) %>% 
  as_tibble()
```

]

.panel[.panel-name[Q2]

```r
# Q2.
ps_data %>% 
  filter(condition == "No Label") %>% 
  arrange(desc(age))
```

```
## # A tibble: 288 x 5
##    subid item   correct   age condition
##    <chr> <chr>    <int> <dbl> <chr>    
##  1 SCH23 faces        0  4.82 No Label 
##  2 SCH23 houses       0  4.82 No Label 
##  3 SCH23 pasta        0  4.82 No Label 
##  4 SCH23 beds         0  4.82 No Label 
##  5 SCH1  faces        0  4.82 No Label 
##  6 SCH1  houses       0  4.82 No Label 
##  7 SCH1  pasta        0  4.82 No Label 
##  8 SCH1  beds         0  4.82 No Label 
##  9 SCH19 faces        0  4.79 No Label 
## 10 SCH19 houses       0  4.79 No Label 
## # … with 278 more rows
```

]

.panel[.panel-name[Q3]

```r
# Q3.
ps_data %>% 
  filter(condition == "No Label" & age >= 3)
```

```
## # A tibble: 192 x 5
##    subid  item   correct   age condition
##    <chr>  <chr>    <int> <dbl> <chr>    
##  1 SCH35  faces        0  3.02 No Label 
##  2 SCH35  houses       0  3.02 No Label 
##  3 SCH35  pasta        0  3.02 No Label 
##  4 SCH35  beds         0  3.02 No Label 
##  5 MSCH40 faces        0  3.02 No Label 
##  6 MSCH40 houses       1  3.02 No Label 
##  7 MSCH40 pasta        0  3.02 No Label 
##  8 MSCH40 beds         1  3.02 No Label 
##  9 SCH34  faces        0  3.06 No Label 
## 10 SCH34  houses       0  3.06 No Label 
## # … with 182 more rows
```

]

.panel[.panel-name[Q4]

.pull-left[

```r
# Q4. 
ps_filtered <- ps_data %>% 
  filter(age >= 3 & age <= 4)

ps_filtered
```

```
## # A tibble: 204 x 5
##    subid item   correct   age condition
##    <chr> <chr>    <int> <dbl> <chr>    
##  1 T10   faces        0  3    Label    
##  2 T10   houses       1  3    Label    
##  3 T10   pasta        1  3    Label    
##  4 T10   beds         1  3    Label    
##  5 T3    faces        1  3.09 Label    
##  6 T3    houses       1  3.09 Label    
##  7 T3    pasta        1  3.09 Label    
##  8 T3    beds         1  3.09 Label    
##  9 T6    faces        1  3.1  Label    
## 10 T6    houses       1  3.1  Label    
## # … with 194 more rows
```
]

.pull-right[

```r
# Q4.
ps_filtered <- ps_data %>% 
  filter(between(age, 3, 4))

ps_filtered
```

---
class: inverse, center, middle

# Manipulating variables <br> (columns)

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `select()`

### Select columns with `select()`

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `select()`

.panelset[

.panel[.panel-name[Arguments]

<code class ='r hljs remark-code'>select(<span style="color:cornflowerblue">.data</span>, <span style="color:red">...</span>)</code>

**.blue[.data]** .blue[= a data frame or tibble]

**.red[. . .]**  .red[= One or more unquoted expressions separated by commas. <br><br> Variable names can be used as if they were positions in the data frame, so expressions like `x:y` can be used to select a range of variables.]
]

.panel[.panel-name[Example 1]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g) %>% 
  select(species:bill_depth_mm)
```

```
## # A tibble: 3 x 4
##   species island bill_length_mm bill_depth_mm
##   <fct>   <fct>           <dbl>         <dbl>
## 1 Gentoo  Biscoe           55.9            17
## 2 Gentoo  Biscoe           55.1            16
## 3 Gentoo  Biscoe           59.6            17
```

]

.panel[.panel-name[Example 2]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g) %>% 
  select(species, starts_with("bill_"))
```

```
## # A tibble: 3 x 3
##   species bill_length_mm bill_depth_mm
##   <fct>            <dbl>         <dbl>
## 1 Gentoo            55.9            17
## 2 Gentoo            55.1            16
## 3 Gentoo            59.6            17
```
]

.panel[.panel-name[Example 3]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g) %>% 
  select(ends_with("_mm"))
```

```
## # A tibble: 3 x 3
##   bill_length_mm bill_depth_mm flipper_length_mm
##            <dbl>         <dbl>             <int>
## 1           55.9            17               228
## 2           55.1            16               230
## 3           59.6            17               230
```
]

]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Selection helpers

Selection helpers work in concert with `select()` to make it easier to select specific groups of variables.

Here are some commonly useful ones

***

`everything()`: Matches all variables.

`last_col()`: Select last variable, possibly with an offset.

`starts_with()`: Starts with a prefix.

`ends_with()`: Ends with a suffix.

`contains()`: Contains a literal string.

.footnote[🔗 https://dplyr.tidyverse.org/reference/dplyr_tidy_select.html#overview-of-selection-features]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Concept Map: `select()`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `mutate()`

### Create (or overwrite) variables with `mutate()`

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `mutate()`

.center[
<img src="images/dplyr_mutate.png" width="75%" />
]

.footnote[Artwork by [@allison_horst](https://twitter.com/allison_horst)]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `mutate()`

.panelset[

.panel[.panel-name[Arguments]

<code class ='r hljs remark-code'>mutate(<span style="color:cornflowerblue">.data</span>, <span style="color:red">...</span>)</code>

**.blue[.data]** .blue[= a data frame or tibble]

**.red[. . .]**  .red[=  Name-value pairs. The name gives the name of the column in the output.]

]

.panel[.panel-name[Example 1]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g) %>% 
  select(starts_with("bill_")) %>% 
  mutate(bill_length_m = bill_length_mm/1000)
```

```
## # A tibble: 3 x 3
##   bill_length_mm bill_depth_mm bill_length_m
##            <dbl>         <dbl>         <dbl>
## 1           55.9            17        0.0559
## 2           55.1            16        0.0551
## 3           59.6            17        0.0596
```

]

.panel[.panel-name[Example 2]

```r
penguins %>% 
  filter(species == "Gentoo" & bill_length_mm > 55) %>% 
  arrange(body_mass_g) %>% 
  select(starts_with("bill_")) %>% 
  mutate(bill_length_mm = as.character(bill_length_mm))
```

```
## # A tibble: 3 x 2
##   bill_length_mm bill_depth_mm
##   <chr>                  <dbl>
## 1 55.9                      17
## 2 55.1                      16
## 3 59.6                      17
```
]

]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Concept Map: `mutate()`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
class: inverse

# Your turn 2

1. In `ps_data`, select only the variables `age` and `condition`.

1. As we did with indexing in base R, you can use the minus sign (`-`) to "de-select" columns. Keep everything in `ps_data` except `subid` and `condition`.

1. Select the columns `correct` and `condition` without naming them, using their positions or de-selecting other variables.

1. Use `mutate()` to convert `condition` from type "character" to type "factor". Then check that you've done this successfully. Hint: what function could you pipe to directly to check this?

---
class: solution

# Solution

.panelset[
.panel[.panel-name[Q1]

```r
ps_data %>% 
  select(age, condition)
```

```
## # A tibble: 588 x 2
##      age condition
##    <dbl> <chr>    
##  1  2    Label    
##  2  2    Label    
##  3  2    Label    
##  4  2    Label    
##  5  2.13 Label    
##  6  2.13 Label    
##  7  2.13 Label    
##  8  2.13 Label    
##  9  2.32 Label    
## 10  2.32 Label    
## # … with 578 more rows
```

]

.panel[.panel-name[Q2]

```r
ps_data %>% 
  select(-c(subid, condition))
```

```
## # A tibble: 588 x 3
##    item   correct   age
##    <chr>    <int> <dbl>
##  1 faces        1  2   
##  2 houses       1  2   
##  3 pasta        0  2   
##  4 beds         0  2   
##  5 beds         0  2.13
##  6 faces        0  2.13
##  7 houses       1  2.13
##  8 pasta        1  2.13
##  9 pasta        0  2.32
## 10 faces        0  2.32
## # … with 578 more rows
```

]

.panel[.panel-name[Q3]

```r
ps_data %>% 
  select(starts_with("c"))
```

```
## # A tibble: 588 x 2
##    correct condition
##      <int> <chr>    
##  1       1 Label    
##  2       1 Label    
##  3       0 Label    
##  4       0 Label    
##  5       0 Label    
##  6       0 Label    
##  7       1 Label    
##  8       1 Label    
##  9       0 Label    
## 10       0 Label    
## # … with 578 more rows
```

]

.panel[.panel-name[Q4]

```r
ps_data %>% 
  mutate(condition = as.factor(condition)) %>% 
  glimpse()
```

```
## Rows: 588
## Columns: 5
## $ subid     <chr> "M22", "M22", "M22", "M22", "T22", "T22", "T22", "T22", "T17", "T17", "T17", "T17", …
## $ item      <chr> "faces", "houses", "pasta", "beds", "beds", "faces", "houses", "pasta", "pasta", "fa…
## $ correct   <int> 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, …
## $ age       <dbl> 2.00, 2.00, 2.00, 2.00, 2.13, 2.13, 2.13, 2.13, 2.32, 2.32, 2.32, 2.32, 2.38, 2.38, …
## $ condition <fct> Label, Label, Label, Label, Label, Label, Label, Label, Label, Label, Label, Label, …
```

]
]

---
class: inverse, center, middle

# Summarizing data

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `summarize()`

`summarize()` reduces your raw data frame into to a smaller summary data frame that only contains the variables resulting from the **summary functions** that you specify within `summarize()`

--
***

Summary functions take vectors as inputs and return single values as outputs

Common examples are `mean()`, `sd()`, `max()`, `min()`, `sum()`, etc...

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `summarize()`

.panelset[

.panel[.panel-name[Arguments]

<code class ='r hljs remark-code'>summarize(<span style="color:cornflowerblue">.data</span>, <span style="color:red">...</span>)</code>

**.blue[.data]** .blue[= a data frame or tibble]

**.red[. . .]**  .red[=  Name-value pairs of summary functions. The name will be the name of the variable in the result.]

]

.panel[.panel-name[Example]

```r
penguins %>% 
  summarize(mean_bill_length = mean(bill_length_mm, na.rm = TRUE),
            max_flipper_length = max(flipper_length_mm, na.rm = TRUE))
```

```
## # A tibble: 1 x 2
##   mean_bill_length max_flipper_length
##              <dbl>              <int>
## 1             43.9                231
```

]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `group_by()`

`group_by()` creates groups based on one or more variables in the data, and this affects any downstream operations -- most commonly, `summarize()`

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `group_by()`

What happens if we combine `group_by()` and `summarize()`?

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# `summarize()`

Let's see a couple examples of how we can combine `group_by()` and `summarize()`

.panelset[

.panel[.panel-name[Example 1]

```r
penguins %>% 
  group_by(species) %>% 
  summarize(mean_bill_length = mean(bill_length_mm, na.rm = TRUE))
```

```
## `summarise()` ungrouping output (override with `.groups` argument)
```

```
## # A tibble: 3 x 2
##   species   mean_bill_length
##   <fct>                <dbl>
## 1 Adelie                38.8
## 2 Chinstrap             48.8
## 3 Gentoo                47.5
```

]

.panel[.panel-name[Example 2]

```r
penguins %>% 
  group_by(species, island) %>% 
  summarize(mean_bill_length = mean(bill_length_mm, na.rm = TRUE))
```

```
## `summarise()` regrouping output by 'species' (override with `.groups` argument)
```

```
## # A tibble: 5 x 3
## # Groups:   species [3]
##   species   island    mean_bill_length
##   <fct>     <fct>                <dbl>
## 1 Adelie    Biscoe                39.0
## 2 Adelie    Dream                 38.5
## 3 Adelie    Torgersen             39.0
## 4 Chinstrap Dream                 48.8
## 5 Gentoo    Biscoe                47.5
```

]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Concept Map: `summarize()`

.footnote[Source: [rstudio/concept-maps](https://github.com/rstudio/concept-maps)]

---
class: inverse

# Your turn 3

1. From `ps_data`, get the total number of correct trials for each subject and call this variable `num_correct`. Hint: you can use the summary function `sum()`.

1. Now get the total number of correct trials (`num_trials`) for each unique combination of `condition` and `item` and arrange the resulting rows by descending order of `num_correct`. Which combination of condition and item had the most/least correct responses?

1. Lastly, calculate the proportion of correct responses for each `condition` and call this variable `prop_correct` (hint: because `correct` is coded as 0 and 1, you can do this by taking the mean of `correct`). In the same pipeline, create a bar plot that shows the differences between the mean proportion of correct responses between the two conditions, and color the bars by condition. What do you notice? (hint: it might be more straightforward to use `geom_col` than `geom_bar`)

---
class: solution

# Solution

.panelset[
.panel[.panel-name[Q1]

```r
# Q1. 
ps_data %>% 
  group_by(subid) %>% 
  summarize(num_correct = sum(correct))
```

```
## `summarise()` ungrouping output (override with `.groups` argument)
```

```
## # A tibble: 147 x 2
##    subid num_correct
##    <chr>       <int>
##  1 C1              4
##  2 C10             2
##  3 C11             3
##  4 C12             3
##  5 C13             2
##  6 C14             3
##  7 C15             4
##  8 C16             2
##  9 C17             2
## 10 C18             3
## # … with 137 more rows
```

]

.panel[.panel-name[Q2]

```r
# Q2. 
ps_data %>% 
  group_by(condition, item) %>% 
  summarize(num_correct = sum(correct)) %>% 
  arrange(desc(num_correct))
```

```
## `summarise()` regrouping output by 'condition' (override with `.groups` argument)
```

```
## # A tibble: 8 x 3
## # Groups:   condition [2]
##   condition item   num_correct
##   <chr>     <chr>        <int>
## 1 Label     beds            60
## 2 Label     pasta           55
## 3 Label     faces           49
## 4 Label     houses          41
## 5 No Label  pasta           18
## 6 No Label  beds            15
## 7 No Label  faces           13
## 8 No Label  houses          12
```

]

.panel[.panel-name[Q3]

```r
Q3. 
ps_data %>% 
  group_by(condition) %>% 
  summarize(prop_correct = mean(correct)) %>% 
  ggplot(aes(x = prop_correct, y = condition, fill = condition)) + 
  geom_col() + 
  coord_flip()
```

]

.panel[.panel-name[Q3 Plot]

![](07-slides_files/figure-html/unnamed-chunk-51-1.png)
]

]

---
background-image: url(images/hex/dplyr.png)
background-position: 90% 5%
background-size: 10%

# Column-wise operations

What if we want to apply `dplyr` verbs across multiple columns **simultaneously**? Check out these slides for more 👇

.footnote[🔗 [columnwise-operations-dplyr.netlify.app](https://columnwise-operations-dplyr.netlify.app/#1)]

---
class: inverse, center, middle
# Q & A

---
class: inverse, center, middle
# Next up...

## Data tidying with `tidyr`

---
class: inverse, center, middle
# Break!