Why Do We Need Reproducibility?

Reproducibility and your code

• Scripts: may make your code reproducible (but not your analysis)

• Commenting and formatting can help!

```{r}
#| eval: false
#|
## load the packages necessary
library(tidyverse)
## read in the data
landmarks.csv <- read_csv("/Users/mattwilliamson/Google Drive/My Drive/TEACHING/Intro_Spatial_Data_R/Data/2023/assignment01/landmarks_ID.csv")

## How many in each feature class
table(landmarks.csv$MTFCC)
```

Reproducible scripts

• Comments explain what the code is doing

• Operations are ordered logically

• Only relevant commands are presented

• Useful object and function names

• Script runs without errors (on your machine and someone else’s)

Toward Efficient Reproducible Analyses

• Scripts can document what you did, but not why you did it!

• Scripts separate your analysis products from your report/manuscript

What is literate programming?

Let us change our traditional attitude to the construction of programs: Instead of imagining that our main task is to instruct a computer what to do, let us concentrate rather on explaining to human beings what we want a computer to do.

What is literate programming?

• Documentation containing code (not vice versa!)

• Direct connection between code and explanation

• Convey meaning to humans rather than telling computer what to do!

• Multiple “scales” possible

Why literate programming?

• Your analysis scripts are computer software

• Integrate math, figures, code, and narrative in one place

• Explaining something helps you learn it

Planning an analysis

Pseudocode and literate programming

• An informal way of writing the ‘logic’ of your program

• Balance between readability and precision

• Avoid syntactic drift

Writing pseudocode

Pseudocode

Start function
Input information
Logical test: if TRUE
  (what to do if TRUE)
else
  (what to do if FALSE)
End function

What is Quarto?

• A multi-language platform for developing reproducible documents

• A ‘lab notebook’ for your analyses

• Allows transparent, reproducible scientific reports and presentations

Key components

1. Metadata and global options: YAML

2. Text, figures, and tables: Markdown and LaTeX

3. Code: knitr (or jupyter if you’re into that sort of thing)

YAML - Yet Another Markup Language

1. Allows you to set (or change) output format

2. Provide options that apply to the entire document

3. Spacing matters!

Formatting Text

• Basic formatting via Markdown

• Fancier options using Divs and spans via Pandoc

• Fenced Divs start and end with ::: (can be any number >3 but must match)

Adding Code Chunks

• Use 3x ``` on each end

• Include the engine {r} (or python or Julia)

• Include options beneath the “fence” using a hashpipe (#|)

Additional considerations

• File locations and Quarto

• Caching for slow operations

• Modularizing code and functional programming