R Workflow Best Practices

class: center, middle, inverse, title-slide

# R Workflow Best Practices
## bmRn CSM: managing code, data, and files
### Martin Frigaard
### 2020-12-13

---

class: center, middle

# R Workflow Best Practices

### Managing your code, data, and files with RStudio

---
class: left, top

## Change RStudio's Default Settings

.pull-left[

#### Click on **Tools** > **Global Options...**

#### - We want to uncheck "*Restore .RData into work space at start up*"

#### - We also want to make sure we change "*Save work space to .Rdata on exit*" to "*Never*"

]

.pull-right[

]

---
class: inverse, center, middle

# Customize RStudio

.pull-left[
### Code
### Console
### Appearance
### Pane Layout
### R Markdown
]

.pull-right[

]

---
class: left, top

# Code Editing

.pull-left[

### - Auto indent?

### - Continue comment lines?

### - Save R scripts before sourcing?

]

.pull-right[

]

---
class: left, top

# Code Display

.pull-left[

### - Margins?

### - Scrolling?

### - Rainbow parentheses?

]

.pull-right[

]

---
class: left, top

# Code Saving

.pull-left[

### - Cursor position?

### - Line endings?

### - Text encoding?

]

.pull-right[

]

---
class: left, top

# Code Completion

.pull-left[

### - Insert parentheses?

### - Insert spaces?

### - Completion delay setting?

]

.pull-right[

]

---
class: left, top

# Code Diagnostics

.pull-left[

### - Check your R Code?

### - Check other languages?

### - How long?

]

.pull-right[

]

---
class: left, top

# Console

.pull-left[

### - Display?

### - Debugging?

### - Other?

]

.pull-right[

]

---
class: left, top

# Appearance

.pull-left[

### - RStudio theme?

### - Zoom?

#### - Also hold `⌘` and press `+` on macOS

#### - Also hold `ctrl` and press `+` on Windows

### - Font?

### - Editor theme?

]

.pull-right[

]

---
class: left, top

## Pane layout

.pull-left[

### - Source?

### - Console?

### Combining pane elements?

#### - Plots, Connections, Build, VCS, Presentation

#### - Files, Packages, Help, Tutorial, Viewer

]

.pull-right[

]

---

# Pane layout view

Standard layout options

---
class: left, top

# Pane layout: add column

#### Two screens?

#### - add a Source column and rearrange the panes

.pull-left[

]

.pull-right[

]

---
class: left, top

# Pane layout: add column view

Now you see **Source**, **Tutorial**, and **Console** panes on a single screen!

---
class: inverse, center, middle

# RStudio Projects

---
class: left, top

# Why RStudio Projects?

### Keep track of all your files with RStudio project files (`.Rproj`).

--
####  Self contained

Using R projects keeps track or your current working directory!

--
#### Project orientated

.Rproj files make bundling and shipping files and folders easier!

#### Avoid removing all the files

???

keep all the files associated with a project together — input data, R scripts, analytic results, figures.

---
class: left, top

# Creating RStudio project in existing folder

#### Click on *'Project: (None)'* > *'New Project'*

---
class: left, top

# Creating RStudio project in existing folder

#### Click on *'Browse* > *'Create Project'*

---
class: left, top

# Creating RStudio projects in new folder

.pull-left[

#### Click on *'Project: (None)'* > *'New Project'*

![Up right tick](img/proj-new-dir.png)
]

.pull-right[

#### Select project type

![Up right tick](img/proj-new-project.png)
]

---
class: left, top

# Creating RStudio projects in new folder

.pull-left[

#### Create new folder name

![](img/proj-new-dir-name.png)

]

.pull-right[

#### Choose parent folder

![](img/proj-new-parent-dir-name.png)
]

---
class: left, top

# Creating RStudio projects in new folder

#### If you have Git installed, select *'Create a git repository'*

---
class: left, top

# Creating RStudio projects in new folder

#### Check for new project name & *Git* pane

---
class: inverse, left, middle

# Folder Structure

- separate raw and cleaned data

- keep documents and code separate

- keep figures separate

- name files appropriately (preferably 2 digit prefix)

- structure is reusable and easy to understand

---
class: left, top

#### Adapted from from '[Good enough practices in scientific computing](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1005510)'

---
class: inverse, center, middle

# Naming things

---
class: left, top

# Naming files*

### File names should be:

.pull-left[

#### 1. human readable -> (makes sense)

#### 2. machine readable -> (regex)

#### 3. sort/order well -> (ISO 8601 date)

]

.pull-right[

##### 2020-10-12-270-301-**central-lab-metrics**.csv

##### 2020-10-12-**270-301**-central-lab-metrics.csv

##### **2020-10-12**-270-301-central-lab-metrics.csv

]

We can perform regular expression searches for files like this:

*Find 270-301 files*

```r
grepl(pattern = "270-301", 
      x = "2020-10-12-270-301-central-lab-metrics.csv")
```

```
## [1] TRUE
```

> *Adapted from [Jenny Byran's slides](https://speakerdeck.com/jennybc/how-to-name-files)

---
class: left, top

# Naming files*

### Also acceptable:

Logical order and underscores `_`

```r
files
```

```
## [1] "01.0-import_270-301_central-lab-metrics.R" 
## [2] "02.0-wrangle_270-301_central-lab-metrics.R"
## [3] "03.0-eda_270-301_central-lab-metrics.R"    
## [4] "04.0-model_270-301_central-lab-metrics.R"
```

```r
stringr::str_split_fixed(string = files, pattern = "_", 3)
```

```
##      [,1]           [,2]      [,3]                   
## [1,] "01.0-import"  "270-301" "central-lab-metrics.R"
## [2,] "02.0-wrangle" "270-301" "central-lab-metrics.R"
## [3,] "03.0-eda"     "270-301" "central-lab-metrics.R"
## [4,] "04.0-model"   "270-301" "central-lab-metrics.R"
```

> *Adapted from [Jenny Byran's slides](https://speakerdeck.com/jennybc/how-to-name-files)

---
class: inverse, center, middle

# File paths

---
class: left, top

# Use relative rather than absolute file paths

### **Absolute paths** are specific to a system

`/project-name/data` -> absolute path in macOS

`\\project-name\\data` -> absolute path in Windows

### **Relative paths** are specific to a folder

`project-name/data` -> relative path in macOS

`project-name\\data` -> relative path in Windows

---
class: left, top

# Or use the `here` package

### The `here::set_here()` function solves a lot of file path problems (*especially if you're not using R projects*)

```r
library(here)
here::set_here(".")
list.files(all.files = TRUE, pattern = "here")
```

```
## [1] ".here"
```

This creates a `.here` file (similar to `.Rproj` files)

As long as the `.here` file stays in the referenced folder, you can include simply include `here::here()` in the top of your code files.

---
class: left, top

# Terminal pane

#### Learn a handful of command-line tools to make life easier

`cd`, `pwd`, `mkdir`, `rm`, `ls`, etc.

#### RStudio comes with a Terminal pane for quick access to the command-line

---
class: left, top

# Getting help

R comes with a *ton* of accessible help files

```r
?read.csv
```

---
class: left, top

## Getting help online

R also help an incredible community! Click on the links below to see some of the common places for Q & A.

### 1) [Dedicated forum on RStudio Community](https://community.rstudio.com/)

### 2) [Questions tagged R on StackOverflow](https://stackoverflow.com/questions/tagged/r)

### 3) [Twitter topics with #rstats hashtag](https://twitter.com/hashtag/rstats)

---
class: left, top

## Asking good questions (reproducible examples)

You'll get better results if you ask a question with a reproducible example. The [`reprex` package](https://github.com/tidyverse/reprex) was designed to help you create one!

```r
install.packages("reprex")
library(reprex)
```

### *Use the RStudio Addin to create a reproducible example from code you've copied onto your clipboard!*

---
class: left, top

# Reprex Addin 1

1. Copy code
2. Select *Addin* > *Render selection*

---
class: left, top

# Reprex Addin 2

1. Copy code
2. Select *Addin* > *Render selection*  
3. Wait for console  
4. Paste reprex

---
class: left, top

# Reprex + datapasta

### To copy + paste actual data in a reproducible example, try `datapasta`!

https://reprex.tidyverse.org/articles/articles/datapasta-reprex.html

---
class: left, top

Learn more about R best practices:

1. [R for Data Science](https://r4ds.had.co.nz/)

2. [Tidyverse](https://www.tidyverse.org/)

3. [RViews Community Blog](https://rviews.rstudio.com/)

---
class: center, top

# THANK YOU!

## Feedback

@mjfrigaard on Twitter and Github

mjfrigaard@gmail.com