index.Rmd

---
title: "Prepare your data and your code for manuscript"
author: "inSileco Team"
output:
  xaringan::moon_reader:
    css: [default, rd.css, rd-font.css, "hygge"]
    lib_dir: assets
    seal: false
    nature:
      highlightStyle: dracula
      countIncrementalSlides: false
      beforeInit: "macros.js"
---


```{r setup, include = FALSE}
source("_setup.R")
htmltools::tagList(
  xaringanExtra::use_clipboard(
    button_text = "<i class=\"fa fa-clipboard\"></i>",
    success_text = "<i class=\"fa fa-check\" style=\"color: #37abc8\"></i>",
  ),
  rmarkdown::html_dependency_font_awesome()
)
xaringanExtra::use_scribble()
```



class: title-slide, middle


## .font140[Data and code ready for your paper]

<br>

.instructors[
  <br><br><br>
  .font120[.authors140[Kevin, Matt & Chris]]
  <br><br>
  .font120[`r format(Sys.time(), '%B %d, %Y')`]
  <br>
]

<br><br><br><br><br><br>

.font140[[`r gh()` dataReady](https://github.com/McCannLab/dataReady/)] 

![deploy workshop](https://github.com/McCannLab/dataReady/workflows/deploy%20workshop/badge.svg)

.instructors[Content under [`r rfa("creative-commons")` `r rfa("creative-commons-zero")`](https://creativecommons.org/licenses/by/4.0/) unless otherwise specified.]





---

class: inverse, center, middle

# Reproducible Research

![:custom_hr]()

## [workshop_reproducibility](https://insilecoinc.github.io/workshop_reproducibility/#1)


---
# Reproducible research (in a nutshell)

--

* Started late 80's 

--

* Not that easy to redo the process: **raw data** `r ar()` **results** `r ar()` **paper**

--

* From 80's to 2010's it get worst (due to an increased number of computational operations)

--

* Reproducible research became more important since 


---

# Reproducible research (in a nutshell)


> We define **reproducibility** to mean **computational reproducibility**—­
obtaining consistent computational results using the same **input data**, computational steps, methods, **code**, and conditions of analysis;


.pull-left[
```{R, echo = FALSE, out.width = "50%", purl = FALSE}
kig("img/cover_rrs.png")
```
]

.pull-right[

`r rfa("file")` National Academies of Sciences, Engineering, and Medicine
(2019). [Reproducibility and Replicability in Science](https://www.nap.edu/catalog/25303/reproducibility-and-replicability-in-science).
]

--

### To do this science need to be **open** at least to some extent: **code** and **data** have to be available.



---
# Reproducible research (in a nutshell)

Make sure that&nbsp;: 

* Data are not a well-chosen subset....
* Code can be run not just on your computer...
* Avoid [excel-related error](https://www.nytimes.com/2013/04/19/opinion/krugman-the-excel-depression.html)
* etc. 

--

### Preparing data and code for submission is an important part of this!



---

class: inverse, center, middle

# Try to be organized

![:custom_hr]()


---
# One well-structured folder! [Download it](https://github.com/inSilecoInc/workshop_reproducibility/raw/main/demo/min_workflow.zip).

--



.pull-left[
```sh
`r emj("file_folder")`
.
├── data
├── ms
├── output
├── src
└── README.md
```
]

--

.pull-right[
```sh
`r emj("file_folder")`
.
├── data
├── ms
├── output
├── R
└── README.md
```
]

<br>

--

`r rfa("blog")` Rich FitzJohn (2013). [Designing projects](https://nicercode.github.io/blog/2013-04-05-projects/)



---
# One well-structured folder!


- `r emj("file_folder")` `R` : R code (`.R` files)

  - functions, scripts ...
  - data retrieval, statistical analysis, figure

<br>

--

- `r emj("file_folder")` `data` :

  - input/raw data
  - processed data
  - data may be archived on line (e.g. Dryad) and may be accessed via a script (see [rdryad](https://docs.ropensci.org/rdryad/))

---
# One well-structured folder!


- `r emj("file_folder")` `output` (`fig` / `tab` / `res`):

  - figures
  - tables

<br>

--

- `r emj("file_folder")` `ms` (`doc`)

  - main manuscript
  - supplementary information (SI)



---
# Follow 4 simple rules!

## 1. Start off raw data 

- avoid editing raw data 
- keep scripts that generate (cleaner) intermediate files
- if you create a lot of simulation, keep the scripts that generate the data.

---
# Follow 4 simple rules!


## 2. The folder must be self contained

  - think of the root of the folder as the root of the system
  - ~~C:\user/data\dat0001.txt~~
  - but this system is not isolated (e.g. remote)
  - does not mean everything will be shared / added in the research compendium

---
# Follow 4 simple rules!

## 3. document massively

  - your datasets (could be a link to relevant documentation)
  - your functions and scripts
  - and the tricks you've used that won't be clear to others (that includes future yourself!)
  - use metadata [EML](https://eml.ecoinformatics.org/)

---
# Follow 4 simple rules!

## 4. Keep track of all versions you've used

  - R version (and other software)
  - Package versions

<br><br>

.font80[`r fl()` Sandve et al. [Ten Simple Rules for Reproducible Computational
Research](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003285)]


---
# One well-structured folder!

### 1. Submit your ms 

### 2. Once your ms is accepted zip the folder without `ms`

### 3. `r ar()` research compendium

--

<br>

```{R, echo = FALSE, out.width = "80%", purl = FALSE}
kig("img/spectrum.png")
```



---

class: inverse, center, middle

# Data & code ready for submission

![:custom_hr]()



---
# What do journals want?

```{R, echo = FALSE, out.width = "50%", purl = FALSE}
kig("img/data_avail.png")
```


.font80[[`r fl()`](https://www.sciencedirect.com/science/article/pii/S0169534715002906) Mislan et al. (2016)]



---
# Data and code availability statements


### Before publication 

> Code and data will made (publicly) available upon acceptance.


### After publication 

> Code and data are available at the following URL (DOI).



---
# Various options 

```{R, echo = FALSE, out.width = "100%", purl = FALSE}
kig("img/repo_table.png")
```

`r fl()` Mislan, K.A.S., Heer, J.M., and White, E.P. (2016). [Elevating The Status of Code in Ecology](https://www.sciencedirect.com/science/article/pii/S0169534715002906). Trends in Ecology & Evolution 31, 4–7.

`r lk()` https://www.nature.com/nature-portfolio/editorial-policies/reporting-standards


---
# Various options 

### Document your data 

* A simple text  
* Using a standard for your metadata ([EML](https://eml.ecoinformatics.org/))
* A data paper 

--

### Document your code 

* Comments 
* Use specific guidelines 
* Create a package
 


---
# Demo/examples

## GitHub + Zenodo 

- https://github.com/McCannLab/NDID

- https://github.com/McCannLab/HomogenFishOntario

## GitHub + Zenodo + Dryad

- https://www.mdpi.com/2304-8158/10/4/717

- https://doi.org/10.5061/dryad.95x69p8jd

- https://zenodo.org/record/4602032#.YHY6RyUpCV4