How to be a pollinatoR

class: title-slide
background-image: url(data:image/png;base64,#images/cover-1-ws.png)
background-size: cover

# How to be a pollinatoR

#### rstudio::conf(2022)

Kiegan Rice  
Weihuang Wong

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Bees

--
.strong[R users have a lot in common with bees.]

.pull-left[
<img src="data:image/png;base64,#images/bee-honeycomb.jpg" title="A bee crawling on honeycomb." alt="A bee crawling on honeycomb." width="75%" />
]

.pull-right[
<img src="data:image/png;base64,#images/rstudio-hexes.jpg" title="R package hex stickers" alt="R package hex stickers" width="95%" />
]

.footnote[
[1] https://commons.wikimedia.org/wiki/File:Western_honey_bee_on_a_honeycomb.jpg  
[2] https://www.rstudio.com/blog/r-and-rstudio-the-interoperability-environment-for-data-analytics/
]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Bees

.strong[Bees are pollinators.]

.pull-left[
<blockquote>
The pollinator population of an area is a great indicator of the overall health of an ecosystem.3 
</blockquote>
]

.pull-right[
<img src="data:image/png;base64,#images/bee-flower.jpg" title="A bee with a flower." alt="A bee with a flower." width="60%" style="display: block; margin: auto;" />
]

.footnote[
[3] https://www.nrcs.usda.gov/wps/portal/nrcs/detail/pa/plantsanimals/?cid=nrcs142p2_018171  
[4] Image: https://www.flickr.com/photos/supersum/5853429721
]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

<h2>
Like bees, R users can enrich their own data ecosystem, with effort and empathy.
</h2>

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Data ecosystem

<!-- 
A few words about who we are, what is this thing we call a data ecosystem, and why we care about it.

Kiegan and I work at NORC, a nonpartisan social research organization. At NORC, our coworkers, partners, and clients use a variety of software: R, of course, but also SAS, Excel, and Stata. We create lots of documentation that accompany our data products, also in a variety of formats, like Word and PDF. We work with folks who are expert coders, beginner coders, non-coders. That's what we mean by an ecosystem: an environment where there are people with a variety of skills, using a variety of software, creating a range of products.

As R users and developers, we play a lot of roles on projects. And just like bees, to complete tasks we need to leave the hive and interact with others. The great thing about R is that other R developers have done a lot of the hard work to help with this interaction, and we'll talk about some of these tools in a second.
-->

---
count: false
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Data ecosystem

---
count: false
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Data ecosystem

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

<!-- 
By enriching the data ecosystem, we mean helping others be better at doing their jobs, regardless of the tools they use or their level of coding expertise. This is what pollinators do: they help everyone else thrive.

Why do we believe so strongly in the role of R users as pollinators?
-->

## strengthen the ecosystem
we succeed or fail together

## make honey 
we do better work when others do better work

## build the hive
build institutional support for R

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### Technical empathy

.strong[Umwelt]
<blockquote>
an Umwelt is... an animal['s]... perceptual world.... Our Umwelt is all we know, and so we easily mistake it for all there is to know.... Humans, however, possess the unique capacity to appreciate the Umwelten of other species....4 
</blockquote>

.footnote[
[5] Ed Yong. "Our Blinding, Blaring World." The Atlantic July/August 2022, pp. 65. 
[6] "Echolocation." http://www.nhc.ed.ac.uk/index.php?page=24.134.166.169.182
]

---
class: primary, inverse
background-image: url(data:image/png;base64,#images/section-1-ws.png)
background-size: cover

# 1. Do no harm

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 1. Do no harm

.pull-left[
<img src="data:image/png;base64,#images/gss-maeduc-missing2.png" title="GSS MAEDUC variable description and summary table" alt="GSS MAEDUC variable description and summary table" width="85%" />

.footnote[[7] 2021 GSS Documentation and Public Use File Codebook (Release 2). https://gss.norc.org/Documents/codebook/GSS%202021%20Codebook.pdf] 
]
--

.right-image[

```r
read_dta(...)
# A tibble: 22 x 2
 maeduc n
 <dbl+lbl> <int>
 1 0 [no formal schooling] 55
 2 1 4
 3 2 7
 ...
22 NA(d) 419
```
]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 1. Do no harm

use {haven} with labelled and tagged_na (foreign::read.dta also has an implementation)

See {haven}'s "Conversion semantics" vignette for more information.

.left-code[

```r
read_dta(...) %>% 
  mutate( 
      var_recoded = labelled(var, 
*     c("Not applicable" = tagged_na("i"),
*       "No answer" = tagged_na("n"),
*       "Dont know" = tagged_na("d")
        )) 
  )
```
]

.right-image[

```r
# A tibble: 24 x 2
 maeduc_recoded n
 <fct> <int>
 1 0 55
 2 1 4
 3 2 7
 ...
22 Dont know 316
23 Not applicable 100
24 No answer 3
```
]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

### 1. Do no harm

Formatting as data: use {openxlsx}, {tidyxl} to retain formatting information

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

## A bit of effort, a bit of empathy.

## Take care with other people's data.

## Do no harm.

---
class: primary, inverse
background-image: url(data:image/png;base64,#images/section-1-ws.png)
background-size: cover

# 2. Make things that just work

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 2. Make things that just work

.pull-left[
Outputting data

```r
# Usual way
write.csv(...)

# Better way
library(foreign)
write.foreign(..., package = "SAS")
```
]

.right-image[
SAS script

<img src="data:image/png;base64,#images/foreign-sas-script.jpg" title="SAS script generated by write.foreign function" alt="SAS script generated by write.foreign function" width="75%" />
]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 2. Make things that just work

--


<img src="data:image/png;base64,#images/renv-example-01.png" title="Email message with text 'Hi Kiegan, looks like there is a hiccup. The server-hosted dashboard is not filtering based on the uploaded file. It is working on my local computer." alt="Email message with text 'Hi Kiegan, looks like there is a hiccup. The server-hosted dashboard is not filtering based on the uploaded file. It is working on my local computer." width="100%" />

--
 
 
<img src="data:image/png;base64,#images/renv-logo.png" title="renv package hex sticker." alt="renv package hex sticker." width="15%" />

.footnote[renv package: https://rstudio.github.io/renv/articles/renv.html]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

## A bit of effort, a bit of empathy. 
## A positive interaction.
## Things just work.

---
class: primary, inverse
background-image: url(data:image/png;base64,#images/section-1-ws.png)
background-size: cover

# 3. Promote growth

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 3. Promote growth

Having newer R users interact with existing scripts and processes.

.pull-left[

<img src="data:image/png;base64,#images/dialog-example-01.png" title="Chat messages. Message 1: 'Can you describe the sequence of lines that you ran?' Message 2: 'So you ran 1-57, then changed overwrite_IU = TRUE, then ran 61-66, then ran 37-57?' Message 3: 'all in the same session'?" alt="Chat messages. Message 1: 'Can you describe the sequence of lines that you ran?' Message 2: 'So you ran 1-57, then changed overwrite_IU = TRUE, then ran 61-66, then ran 37-57?' Message 3: 'all in the same session'?" width="80%" style="display: block; margin: auto;" />
]

.pull-right[
{svDialogs}
<img src="data:image/png;base64,#images/dialog-box.jpg" title="A dialog box pop-up in R asking the user if they would like to include NA values for a particular value in an analysis. The user can select 'yes' or 'no' in the pop-up box." alt="A dialog box pop-up in R asking the user if they would like to include NA values for a particular value in an analysis. The user can select 'yes' or 'no' in the pop-up box." width="70%" style="display: block; margin: auto;" />

]

.footnote[svDialogs package: https://cran.r-project.org/web/packages/svDialogs/svDialogs.pdf]

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 3. Promote growth

---
count: false
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 3. Promote growth

---
count: false
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### 3. Promote growth

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]

## A bit of empathy, a bit more effort. 
## A learning opportunity. 
## Promote growth.

---
.slide-header[HOW TO BE A POLLINATO.strong[R]]
### ...how can you pollinate your own data workflows using R?

.strong[Can you...]

- do no harm?  
  - make things that just work?  
  - promote growth?

---
class: inverse
background-image: url(data:image/png;base64,#images/thankyou-1-ws.png)
background-size: cover

.contact[
.strong[Kiegan Rice]  
Statistician  
rice-kiegan@norc.org  
@kieganer

.strong[Weihuang Wong]  
Senior Research Methodologist  
wong-weihuang@norc.org  
@weihuang
]