day3.Rmd

---
title: "Intro to R 2025 day 3"
author: "Matt Cannon"
date: '2025-04-02'
output:
  html_document:
    code_folding: hide
    toc: true
    toc_depth: 2
---

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

Load libraries
```{r}
library(tidyverse)
```

# Day 3

## General stuff

### Common error messages
```{r, eval=FALSE}
mean(bob)
# Error in mean(bob) : object 'bob' not found
```
R is looking for a variable named bob, but it doesn't exist
-   Do you need to make it?
-   Check spelling/capitalization
-   Did you forget to put quotes around a character string perhaps?

```{r, eval=FALSE}
arbleGarble(mtcars)
# Error in arbleGarble(mtcars) : could not find function "arbleGarble"
```
R is trying to use the function `arbleGarble()`, but it doesn't exist
-   Either load in the library that has it
    -   (or use `coolLib::arbleGarble()`)
-   Check spelling/capitalization

```{r, eval=FALSE}
hist(mtcars$mpg, n = 20
# +
((((((1))))))
```
R thinks you have unfinished business here
-   Hit escape to cancel the command
-   Most likely you're missing a parenthesis or a quotation mark somewhere
-   Rainbow brackets for the win!

### googling error messages

### googling "R how do I…"

### ChatGPT

### Be careful using code from the web/AI!
- Make sure you understand what it does
- You'll learn a lot faster if you read the documentation and try to understand what's happening
- You can still use code from the web, but try to understand it first!

---

## Functions for today's activity

### cbind(), rbind()
- attach two data.frames either side to side or top to bottom
- `rbind()` binds rows together  and will arrange column names
    - Stacks data on top of each other
- `cbind()` binds columns together just puts things together without checking order
    - Put data side by side
    - Generally a bad idea to cbind() unless you're very careful!

```{r}
efficient_cars <-
  mtcars %>%
  filter(mpg > 30)

gas_guzzlers <-
  mtcars %>%
  filter(mpg < 14.5)

rbind(efficient_cars, gas_guzzlers)

# This is total nonsense, but R does it anyways
cbind(efficient_cars, gas_guzzlers)
```

### merge()
- merge() will attach two data.frames side to side
- It uses shared columns to match data into the right rows
- By default, does an inner join
    - Keeps only rows that have matching data in both data.frames
- Use the arguments `all`, `all.x` or `all.y` to specify other join types

https://www.datasciencemadesimple.com/wp-content/uploads/2017/06/merge-in-R-2.jpg?ezimgfmt=rs:535x142/rscb1/ng:webp/ngcb1

band_members and band_instruments are two pre-loaded data.frames
```{r}
band_members
band_instruments
merge(band_members, band_instruments)
merge(band_members, band_instruments, all = TRUE)
```

### mutate()
- Add a new column or change an existing one
- Takes in a data.frame and a new column name and value
- Can use other columns in the data.frame to calculate the new column

```{r}
km_per_mi <- 1.60934

mtcars %>%
  mutate(kpg = mpg * km_per_mi) %>%
  head()

mtcars %>%
  mutate(mpg = mpg * km_per_mi)
```

### summarize() and group_by()
- `group_by()` silently divides up rows by categories
- `summarize()` summarizes data within groups in a data.frame
- Makes a new column and drops all columns not in group_by() or created by summarize()
- `ungroup()` removes the grouping

starwars is another pre-loaded data.frame play dataset
```{r}
mtcars %>%
  group_by(cyl) %>%
  summarize(mean_mpg = mean(mpg)) %>%
  ungroup()

starwars %>%
  group_by(homeworld, species) %>%
  summarize(
    mean_height = mean(height),
    max_mass = max(mass)
  )
```

### t()
- Rotate a dataframe (or matrix) 90 degrees
- A matrix is a 2D array of data similar to a data.frame
    - All the data inside has to be the same type
- `t()` also turns the data into a matrix, so we re-convert it using as.data.frame()
    - You can often convert variable types using `as.the_thing_I_want()`

```{r}
rotated_cars <-
  t(mtcars) %>%
  as.data.frame()
```

### pivot_longer()
- wide data vs long data
    - Wide data has multiple columns that hold the same type of data
        - For instance, a column for each year
    - Long data has a single column for each data category and other columns for metadata
        - For instance, a column for year and a column for value
- A lot of tidyverse functions work best with long data
- pivot_longer() takes wide data and makes it long
- "tidy" data - https://www.youtube.com/watch?v=K-ss_ag2k9E explanation starts ~ 6 min in
- https://epirhandbook.com/en/images/pivoting/pivot_longer_new.png
- https://fromthebottomoftheheap.net/2019/10/25/pivoting-tidily/

- For pivot_longer()
    - The cols argument is the columns that are in wide format
        - here it is everything but "religion", all the income columns
    - The names_to argument is the name of the column that will hold the column names
        - We want our new column to be named "income"
    - The values_to argument is the name of the column that will hold the values
        - We want our new column to be named "count"
- For pivot_wider()
    - The names_from argument is the name of the column that holds the column names
        - We want to use the "income" column
        - The values in this column will become the new column names
    - The values_from argument is the name of the column that holds the values
        - We want to use the "count" column
        - These values will be put into the new columns beneath the corresponding column name

```{r}
# wide data
relig_income

# long data
long_data <-
  relig_income %>%
  pivot_longer(
    cols = !religion,
    names_to = "income",
    values_to = "count"
  )

# Make it wider
wide_again <-
  long_data %>%
  pivot_wider(
    names_from = "income",
    values_from = "count"
  )
```

### Write a table out to a text file
```{r}
write.table(
  wide_again,
  file = "exampleOutput.txt",
  quote = FALSE,
  sep = "\t",
  row.names = FALSE
)

write_tsv(
  wide_again,
  file = "exampleOutputAlso.txt"
)
```

## Activity
- Use patientGroups.txt and exercise.txt from the materials folder
    - patientGroups is patient # and who received treatment
    - Exercise is the how many minutes each patient exercised across five days
- Combine the datasets into a single data.frame
- Pivot the data to long form
    - Columns: patient, day, exercise_min, glucose, trt_group
- Save the pivoted data.frame to a text file with write.table()
- Make a new column by multiplying glucose by 1000
- Calculate the average daily minutes of exercise per patient
- If you're super fast:
    - Help someone who is struggling
    - Plot glucose levels for each group (treated/control)
    - Test if they're statistically different
    - Plot daily exercise minutes per group
    - Test if exercise minutes was statistically different between groups

### Read in patientGroups.txt and exercise.txt
patientGroups.txt is:

-   patient #
-   treatment groups
-   final blood glucose measurements

exercise.txt is:

-   patient #
-   how many minutes each patient exercised across five days

```{r}

```

### Combine the datasets into a single data.frame
```{r}

```

### Pivot the data to long form
Columns should be patient, day, exercise_min, glucose
```{r}

```

### Save the pivoted data.frame to a text file with write.table() or write_tsv()
```{r}

```

### Make a new column where glucose is multiplied by 1000
```{r}

```

### Calculate the average daily minutes of exercise per patient
```{r}

```

## If you're super fast

### Plot glucose levels for each group (treated/control)
```{r}

```

### Test if glucose levels are statistically different between groups
```{r}

```

### Plot daily exercise minutes per group
```{r}

```

### Test if exercise minutes is statistically different between groups
```{r}

```