02-data-manipulation-solutions.Rmd

---
title: "Data Manipulation Exercises"
author: "Joscelin Rocha Hidalgo"
output: 
    html_document:
        css: style.css
        toc: true
        toc_depth: 1
        toc_float: true
        df_print: paged
---


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

# Load Packages

Load the `tidyverse` package and the `janitor` package.

```{r}
library(tidyverse)
library(janitor)
```

# Import data: chds6162_data

Import the chds6162_data to a dataframe called `data`. 

```{r}
data <- read_csv("data/chds6162_data.csv")
```


# Select

![](images/select.png)
Use `select` to show just the `marital` variable.

```{r}
data %>%
  select(marital)
```

`select` for multiple variables:

Use `select` to show `marital` and `ed` variables.

```{r}
data %>%
  select(marital, ed)
```

Use  `select`  for a range of columns. 

`select` all the variables from `wt` to the end.

```{r}
# option one:

data %>%
  select(wt:number)

# option 2:

data %>%
  select(wt:last_col())
```

Drop the `race` variable.

```{r}
data %>%
  select(-race)
```

Drop the variables that belonged to the father from `drace` to `dwt`

```{r}

data %>%
  select(-(drace:dwt))
```


# Filter
![](images/filter.png)
![art by @allison_horst](https://raw.githubusercontent.com/allisonhorst/stats-illustrations/master/rstats-artwork/dplyr_filter.jpg)
art by @allison_horst

We use `filter` to choose a subset of cases.

Use `filter` to keep only respondents who are divorced (`marital` = 3). 

```{r}
data %>%
  filter(marital == 3)
```

Use `filter` to keep only respondents who are **not** divorced. 

```{r}
data %>%
  filter(marital != 3)
```

Use `%in%` within the `filter` function to keep only those whose mothers' race was reported as white. In this dataset, values from 0 to 5 represent "white" under the `race`variable.

```{r}
data %>%
  filter(race %in% c(0:5))
```

Create a chain that keeps only those are college grads (value of 5). Then, `filter` to keep only those who are not married (value of 5).

```{r}
data %>%
  filter(ed == 5) %>%
  filter(marital != 5)
```

#?

# Mutate

![](images/mutate.png)
![art by @allison_horst](images/dplyr_mutate.png)
art by @allison_horst


Create a **new variable with a specific value**

Create a new variable called `country` and fill that variable with "US".

```{r}
data %>%
  mutate(country = "US")
```


Create a **new variable based on other variables**

Create a new variable called `gestation_weeks` and calculate gestation length in weeks rather than days (try rounding this number to only 2 decimals). Remember that the `gestation` variable is a measure of the **gestation length** in days. Then `select` both variables. 

```{r}
data %>%
  mutate(gestation_weeks = round(gestation / 7, 2)) %>%
  select(gestation, gestation_weeks)
```

Change an **existing variable**

Convert the `dwt` variable to kilos by dividing by 2.205 (it's in pounds now). Then, use `select` to show only the `dwt` variable.

*Try rounding to 2 decimals

```{r}
data %>%
  mutate(dwt = dwt / 2.205) %>%
  select(dwt)

# Rounding it

data %>%
  mutate(dwt = round(dwt / 2.205, 2)) %>%
  select(dwt)
```

# Case_when

![art by @allison_horst](images/dplyr_case_when.png)
art by @allison_horst

Using mutate and case_when, let's create a variable called `age_group`. In this variable we will use the mom's age variable and recode it so all moms whose age are now "20s", "30s", and so on. Then select the `id` and the new variable

*tip= Wanna know the range of your sample age? Use `range(dataframe$variable,na.rm = TRUE)`

** Use `%in%` (you just learned it 😉)

```{r}

range(data$age, na.rm = TRUE)

data %>%
  mutate(age_group = case_when(
    age %in% 10:19 ~ "10s",
    age %in% 20:29 ~ "20s",
    age %in% 30:39 ~ "30s",
    age %in% 40:49 ~ "40s",
    age %in% 50:59 ~ "50s"
  ))
```


# Summarize or Summarise

![](images/summarize.png)

Get the mean, median, min, and max weight for the fathers `dwt`.

```{r}
data %>%
  summarize(
    mean_dad_weight = mean(dwt, na.rm = TRUE),
    median_dad_weight = median(dwt, na.rm = TRUE),
    min_dad_weight = min(dwt, na.rm = TRUE),
    max_dad_weight = max(dwt, na.rm = TRUE)
  )
```

#?

# Group_by

![](images/group-by.png)
![art by @allison_horst](images/group_by_ungroup.png)
art by @allison_horst

Calculate the mean weight for fathers based on their education levels (`ded`variable has some NAs, don't forget to remove them!)
 with `na.rm=`)
 
```{r}
data %>%
  group_by(ded) %>%
  drop_na(ded) %>%
  summarize(dad_weight = mean(dwt,
    na.rm = TRUE
  ))
```

# Across
![art by @allison_horst](images/dplyr_across.png)
art by @allison_horst

We messed up and we found out that out collaborators need the mean of all the variables for dad by their race category (dads'). 

Use `across` to get them all!
* Don't forget to remove NAs otherwise it will yell at you!

```{r}
data %>%
  drop_na(drace) %>%
  group_by(drace) %>%
  summarize(across(dage:dwt, mean, na.rm = TRUE))
```

#?

# Big exercise

A new data frame called `mothers_below30` that:

1. Uses `filter` to only include those age younger than 30
2. Uses `mutate` to transform the `gestation` variable into weeks rather than days
3. Uses `group_by` to create mom education groups
4. Uses `summarise` to calculate a new variable called `mean_gestation_w`
5. Make sure to export it as a csv!


```{r}
mothers_below30 <- data %>%
  filter(age < 30) %>%
  mutate(gestation_w = gestation / 7) %>%
  group_by(ed) %>%
  summarize(mean_gestation_w = mean(gestation_w, na.rm = TRUE))
write_csv(mothers_below30, "exports/moms_under30.csv")
```


+ Before you export it as a CSV, move one of the variables to a different position using "relocate"

With Bonus:

```{r}
mothers_below30 <- data %>%
  filter(age < 30) %>%
  mutate(gestation_w = gestation / 7) %>%
  group_by(ed) %>%
  summarize(mean_gestation_w = mean(gestation_w, na.rm = TRUE)) %>%
  relocate(mean_gestation_w, .after = ed)
write_csv(mothers_below30, "exports/moms_under30.csv")
```