Geospatial Analysis with R

class: center, middle, inverse, title-slide

.title[
# Geospatial Analysis with R
]
.subtitle[
## Class 4
]

---

---

```r
data("chirps")
library(rasterVis)
library(lubridate)
library(RColorBrewer)
library(raster)
zam <- getData("GADM", country = "ZMB", level = 1)

dates <- seq(as_date("2016-10-25"), as_date("2016-11-21"), by = "day")
dates <- data.frame(dates, w = unlist(lapply(1:4, function(x) rep(x, 7))))
rfweek <- stack(lapply(unique(dates$w), function(x) {
  w <- which(dates$w == x)
  calc(chirps[[x]], sum)
}))
wk <- as_date(sapply(unique(dates$w), function(x) dates[dates$w == x, 1][1]))
names(rfweek) <- wk

mapTheme <- rasterTheme(region = brewer.pal(9, "Blues"))
pstrip <- list(cex = 1, col = "white")
png("figures/weekly-rf.png", height = 5, width = 5, res = 300,
    units = "in", bg = "transparent")
p <- levelplot(rfweek, scales=list(draw = FALSE), par.settings = mapTheme,
               names.attr = paste("Week of", wk), axes = FALSE, xlab = "",
               ylab = "", par.strip.text = pstrip, 
               main = list("Weekly Rainfall (mm)", col = "White"), 
               colorkey = list(axis.line = list(col = "white"), 
                               axis.text = list(col = "white")))
p2 <- p + latticeExtra::layer(sp.polygons(zam, col = "grey80", lwd = 0.7))
p2
dev.off()
```

---
# Today

- A review of where we are so far
- Rmarkdown demo
- Picking up from last class
  - Projects and package set-up
  - `git` / GitHub
      - syncing, commit, push, pull, branching, merging

---
# So far...

- We should know:
  - Key concepts/tools of reproducibility and why we use them
  - How to set up R package project with `git` VCS
      - More on package structures: see [here](https://r-pkgs.org/package-structure-state.html)
  - Where your library lives
  - *What the key ingredients in a package are*
  - *How to keep project synced between local and remote repos*
  - *How to document functions*

- What we might not know (about packages):
  - data folder and lazy loads
  - inst folder and how to get at it

---

## Git/GitHub

<div class="figure" style="text-align: center">
<img src="https://kevintshoemaker.github.io/StatsChats/GIT1.png" alt="kevintshoemaker.github.io/StatsChats/GIT_tutorial" width="70%" />
<p class="caption">kevintshoemaker.github.io/StatsChats/GIT_tutorial</p>
</div>
---
<div class="figure" style="text-align: center">
<img src="https://kevintshoemaker.github.io/StatsChats/GIT2.png" alt="kevintshoemaker.github.io/StatsChats/GIT_tutorial" width="70%" />
<p class="caption">kevintshoemaker.github.io/StatsChats/GIT_tutorial</p>
</div>
---
<div class="figure" style="text-align: center">
<img src="https://i.stack.imgur.com/yof9f.png" alt="stackoverflow.com/questions/7212740/why-git-is-called-a-distributed-source-control-system" width="70%" />
<p class="caption">stackoverflow.com/questions/7212740/why-git-is-called-a-distributed-source-control-system</p>
</div>
---
## Our Branching Model

---
## Merging

---
## Git exercises (1)
- Open your personal R repo (e.g. `mjc346` in RStudio)
- Copy and paste the text in your Description file to a separate file. 
- Create a new branch, `test_merge`. You can create this branch in the Git tab, or in Terminal, using `git checkout -b test_merge`
- With the new branch active, create a new file `test_function` within the R folder. Create a simple function that takes a number `x` as an input, and prints "`x` is my favorite number!". Save and commit this file in the new branch.
- Switch back to the `main` branch. You should see a pop-up message. What happens to the `test_function.R` file? Can you see it in your file explorer?
- merge the new file to the main branch using the command `git merge test_merge` in Terminal. What happens?
- The `merge` command automatically merges and commits changes to the main branch

---
## Git exercises (2)
- In the `test_merge` branch, open the Description file. Update the version to 0.0.2. Save and commit.
- Switch to the `main` branch. Add a second author in the Author line. Save and commit.
- Now test merging the change from `test_merge` to `main` using `git merge test_merge`. Does this work?

---
## Git exercises (3)
- In the `test_merge` branch, add a third fictional author. Save and commit.
- Switch to the `main` branch. Again add a third author, but use a different author than in the `test_merge` branch.
- Now test merging the change from `test_merge` to `main` using `git merge test_merge`. Does this work?
- Open the Description file. What does it look like?
- After you are done, delete the branch using `git branch -D test_merge`
- Update the Description file in the `main` branch by pasting the text you saved. Save and commit. 
---

# Tips and Tricks

- Tab completion and shortcuts
- Reusing code
- Code syntax
- [Posting guide](https://www.r-project.org/posting-guide.html)

---

## Knowing how to get help as a skillset

- Slack posting guide
- Getting help via the search engine
- (Eventually) posting to listserves

---
## Search Engine Science

- Sometimes you just need the error message
<img src="figures/class3_4.jpeg" width="90%" style="display: block; margin: auto;" />

---
## Search Engine Science

- Sometimes you need to search
  ```
  fatal: unable to access 'https://github.com/agroimpacts/xyz346.git/': 
  error setting certificate verify locations:
   CAfile: C:/Users/xyz/Desktop/ADP/RStudio/xyz346/Git/mingw64/ssl/
   certs/ca-bundle.crt
   CApath: none
  ```
- How you search matters

---

---
## Listserves
<img src="figures/class3_3.png" width="90%" style="display: block; margin: auto;" />
---

## Package functions
- `install.packages` : installs binary packages from CRAN
- `devtools::install_github`: installs from Github
-  `devtools::install("C:/Users/micha/Documents/geospaar/")` to install locally

---

## Data in packages

- packages often include example data
- Lazy loading data only loads when used
- Lazy loaded data in data/ folder
  - formats: .R, .rda, .RData, .tab, .txt, .csv
- Non lazy loads (raw data) in inst/extdata

---

```r
ls()
```

```
## character(0)
```

```r
data("farmers_env", package = "geospaar")
ls()
```

```
## [1] "farmers_env"
```

```r
farmers_env
```

```
## Simple feature collection with 793 features and 3 fields
## Geometry type: POINT
## Dimension:     XY
## Bounding box:  xmin: 24.777 ymin: -18.222 xmax: 33.332 ymax: -8.997
## CRS:           NA
## # A tibble: 793 × 4
##    uuid             geometry  rain district
##  * <chr>             <POINT> <dbl>    <dbl>
##  1 009a8424 (27.256 -16.926)  66.9       11
##  2 00df166f (26.942 -16.504)  72.0       11
##  3 02671a00 (27.254 -16.914)  66.9       11
##  4 03f4dcca (27.237 -16.733)  69.8       11
##  5 042cf7b3 (27.138 -16.807)  75.1       11
##  6 05618404 (26.875 -16.611)  66.1       11
##  7 064beba0 (26.752 -16.862)  58.4       20
##  8 083a46a2 (26.977 -16.765)  70.4       11
##  9 08eb2224 (26.912 -16.248)  54.0       59
## 10 0ab761d6  (27.113 -16.95)  72.4       11
## # … with 783 more rows
```

```r
rm(list = ls())
ls()
```

```
## character(0)
```

---

```r
library(geospaar)
ls()
```

```
## character(0)
```

```r
farmers_env
```

```r
ls()
```

```
## character(0)
```

---
## Raw data in inst/extdata

---

```r
system.file("extdata", package = "geospaar")
```

```
## [1] "/Library/Frameworks/R.framework/Versions/4.2/Resources/library/geospaar/extdata"
```

```r
dir(system.file("extdata", package = "geospaar"))
```

```
##  [1] "cdf_corn.csv"          "chilanga_farmer"      
##  [3] "chirps.tif"            "districts.geojson"    
##  [5] "EJ_POLY.cpg"           "EJ_POLY.dbf"          
##  [7] "EJ_POLY.prj"           "EJ_POLY.sbn"          
##  [9] "EJ_POLY.sbx"           "EJ_POLY.shp"          
## [11] "EJ_POLY.shp.xml"       "EJ_POLY.shx"          
## [13] "FAOSTAT_maize.csv"     "FAOSTAT_sorghum.csv"  
## [15] "FAOSTAT_wheat.csv"     "farmer_spatial.csv"   
## [17] "filt_data.rda"         "roads.geojson"        
## [19] "train_reference.csv"   "westluse.rdc"         
## [21] "westluse.rst"          "westroad.vct"         
## [23] "westroad.vdc"          "whittier_down_d_h.rda"
## [25] "whittier_mid_d_h.rda"  "whittier_up_d_h.rda"
```

```r
f <- system.file("extdata", "farmer_spatial.csv", package = "geospaar")
head(read.csv(f))
```

```
##       uuid      x       y       date season rained planted
## 1 009a8424 27.256 -16.926 2015-11-15      1      0       0
## 2 00df166f 26.942 -16.504 2015-11-15      1      0       0
## 3 02671a00 27.254 -16.914 2015-11-15      1      0       0
## 4 03f4dcca 27.237 -16.733 2015-11-15      1      0       0
## 5 042cf7b3 27.138 -16.807 2015-11-15      1      0       0
## 6 05618404 26.875 -16.611 2015-11-15      1      0       0
```

---
## Other package folders
- See [Chap 9 of "R Packages"](https://r-pkgs.org/misc.html) for more examples of R packages folders.

---
# A look at RMarkdown

Chunk options

[Rmarkdown demo](rmarkdown_demo.html) by Lei Song

---
## RMarkdown exercises
- In your personal R project, Create new RMarkdown file using `usethis::use_vignette(name = "example_vignette")`
- Create chunks for the following:
  - Load the `geospaar` package
  - Load the `chirps` data and print its structure using `str(chirps)`
  - Plot `chirps` data using `raster::plot(chirps)`. Above this chunk, include a text description for the CHIRPS data using different text formatting (bold, italics, etc). You can see information on the `CHIRPS` data in the `man` folder.
  - Include a chunk with an obvious error, like `5 + "string"`
  - Try knitting this RMD using different [chunk options](https://rmarkdown.rstudio.com/lesson-3.html)