replconf2024.qmd

---
title: RES Replicator Conference 2024
# title-slide-attributes:
#   data-background-image: img/RES-logo_WHITE copy.png
#   data-background-size: 15%
#   data-background-position: 60% 52%
format: 
  revealjs:
    theme: _extensions/grantmcdermott/clean/clean.scss
    chalkboard: false
    logo: img/RES-logo_WHITE.png
    footer: "[RES Replicator Conference 2024](https://RES-reproducibility.github.io/replconf2024/)"
    incremental: false
    code-line-numbers: false
    highlight-style: github
    slide-number: true
author: 
  - name: Florian Oswald
    orcid: 0000-0001-7737-2038
    email: florian.oswald@sciencespo.fr
    affiliations: SciencesPo Paris, RES Data Editor
subtitle: "[First *ever* RES Replicator Conference ✌️](https://RES-reproducibility.github.io/replconf2024/)"
date: last-modified
date-format: "D MMMM, YYYY"
---

## Agenda


1. How to write a report? News on the `README`.
2. How to deal with confidential data?
3. `STATA` versioning issues.
4. `R` versioning issues.
5. `nuvolos` rules.
6. `docker` workflow.
7. Collecting tips and tricks.

::: {.fragment}


8. 🍻 🥳 🍾

:::

## How to write a Report?

* New guidelines on [EJ workflow](https://res-reproducibility.github.io/EJ-Workflow/#step-4.-replicators) as well as on [template repo](https://github.com/RES-Reproducibility/EJ-report-template)
* `README` now must have a _statement about rights_.

![](img/rights.png)


## How to deal with Confidential Data?

::: {.callout-note}

# Lookout for keywords 🧐

What the replicator should pick up on immediately: the words _confidential_, _restricted access_, _not available_ etc.

:::

**Layered Approach if Data Cannot be Shared:**

::: {.incremental}
1. If we have instructions for direct access, we try (time limit: 30 mins)
2. If not, try to get access to authors/data provider's machine (i.e. their _screen_)
3. If not, data provider may certify results for us.
4. If not, must provide simulated version of data.
:::

## `STATA` Versioning Issues

1. Must be reproducible _with version of author_.
2. Complicated relationship between `STATA` version and package versions.
3. Best practice: [Julian Reif's Guidelines]([insert link here](https://julianreif.com/guide/#libraries))
4. Case Study on [blog post](https://ejdataeditor.github.io/posts/20240505-stataversions/)
  
## `R` Versioning Issues

1. Where are `R` packages installed?
2. Does copying compiled package of author into your `.libPaths()` work?
3. Why is `R` more complicated than `STATA` here?
4. We require full list (with versions) of all `R` packages. How can you guarantee this environment on your system? [`renv`](https://rstudio.github.io/renv/articles/renv.html)


## `nuvolos` Rules


::: {.incremental}
1. You are allowed to create new instances. However,
2. You **must invite me** onto this instance! I need to be able to look at all instances.
3. Configure all your apps to start in _shared mode_.
4. Be careful with storage. Your instance has only 50GB. We can attach a large file storage system on demand.
:::

## `docker` workflow

::: {.incremental}

1. What is [`docker link`](https://www.docker.com)?
2. What do you do if results require `STATA 15`? 👉 [docker stata 15](https://github.com/AEADataEditor/docker-stata)
3. What do you do if results require `INTEL fortran 2023 compiler, MKL, Mosek 1.2, Gurobi 3.4, R 3.2, STATA 16 MP, python 3.9` ? ☠️ 
4. 👉 [docker container here](https://github.com/AEADataEditor/docker-aer-2023-0700)
5. I am trying to move all operations there. tbc.
:::

## Run `STATA` in a `docker` 

1. start docker
2. execute this script

```bash
# save this in a file: mydocker.sh
# define some variables First
TAG=2023-06-13
VERSION=16
MYHUBID=dataeditors
MYIMG=stata$VERSION
STATALIC="$(find $HOME/Dropbox/licenses -name stata.lic | tail -1)"
echo "using license at $STATALIC"
echo "STATA version $VERSION, docker IMG tag $TAG"


docker run -it --rm \
  -v "$STATALIC":/usr/local/stata/stata.lic \
  --name="dock_stata$VERSION" \
  --mount type=bind,source="$(pwd)",target=/project \
  $MYHUBID/$MYIMG:$TAG

# run by saying ./mydocker.sh on your command line
```

## Collecting and Documenting Tips and Tricks

* replicator resources on website
* wiki on EJ-Workflow

## Resources

Here are a few other resources which might be helpful:

* [Workshop Material on Reproducibility at CEPII 2023](https://res-reproducibility.github.io/hitchhiker/)
* Hidden page on DE website: [Replicator Resources](https://ejdataeditor.github.io/replicator-resources)
* AEA data editor [github org](https://github.com/AEADataEditor/)
* ES resources?

# Questions? {background-color="#40666e"}




# End 🍻 {background-color="#40666e"}