Geospatial Visualization using R

class: center, middle, inverse, title-slide

# Geospatial Visualization using R
## Part 4: Spatial Data Operations
### Bhaskar V. Karambelkar
### 2017/07/04

---

# Selectiing features/fields `sp` objects

```r
dim(meuse@data)
```

```
## [1] 155  12
```

```r
colnames(meuse@data)
```

```
##  [1] "cadmium" "copper"  "lead"    "zinc"    "elev"    "dist"    "om"     
##  [8] "ffreq"   "soil"    "lime"    "landuse" "dist.m"
```

```r
print(meuse[9:10, c('lead','zinc')])
```

```
##         coordinates lead zinc
## 9  (181060, 333231)  133  347
## 10 (181232, 333168)   80  183
```

```r
nrow(meuse[meuse$lead>quantile(meuse$lead,.75),])
```

```
## [1] 38
```

---
# Other Operations on `sp` objects

- `[[`: Data for single column `plot(hist(meuse[['zinc']]))`

- `[[<-`: Replace values of single column `meuse[['lead']] <- rnorm(nrow(meuse))`

- `proj4String` / `proj4String<-`: Returns / sets (w/o transforming) the CRS.

- `spTransform(meuse, CRS("+init=epsg:4326"))`: Convert from one CRS to another.

- `geometry(meuse)`: Strip the fields and return the bare geometry.

- `over(points, polygons, returnList=TRUE)`: Returns a list indicating which polygons contain each point in `points` data-set.

- `rgeos` / `gdistance` / `geosphere` / `maptools`: Packages contain many geometry and feature related operations.

---

# Example

```r
par(mar=c(0,0,0,0))
plot(scot_BNG, col='#aaaaaa', border=0)
*plot(rgeos::gCentroid(scot_BNG, byid=TRUE), add=TRUE, pch=20) 
```

---

# Operations on `sf` Objects

- `sp` => `sf`

```r
suppressPackageStartupMessages(library(sf))
*meuse_sf <- st_as_sf(meuse) 
as.data.frame(meuse_sf[1:2,c('lead', 'zinc', 'geometry')])
```

```
##   lead zinc             geometry
## 1  299 1022 POINT(181072 333611)
## 2  277 1141 POINT(181025 333558)
```

- `st_coordinates(sf)`: Extract the X/Y coordinates.

- `st_geometry(sf)`: Just the geometry, the `sfc` column.

- `st_as_text(sfg)` / `st_as_binary(sfg)` convert to WKT and WKB resp.

- `st_as_sfc()` for converting a foreign object into a `sfc`.

- Check out `methods(class='sfc')` for many more methods.1

.footnote[1 - Best place to start are the `sf` vignettes [here](https://cran.r-project.org/web/packages/sf/vignettes/sf1.html) & [here](https://cran.r-project.org/web/packages/sf/vignettes/sf3.html).]

---

# Example

```r
*usa_sfg <- st_transform(st_geometry(albersusa::usa_sf()), albersusa::us_aeqd_proj) 
*usa_c <- st_centroid(usa_sfg) 
*usa_50 <- (usa_sfg - usa_c) * .50 + usa_c 
par(mar=c(0,0,0,0));plot(usa_sfg)
plot(usa_50, add=TRUE, col='#FFA50077', border=0)
plot(usa_c, add=TRUE, pch=20)
```

---

# Operations on `raster` Objects

```r
suppressPackageStartupMessages(library(raster))
r <- raster(system.file("external/test.grd", package="raster"))
inMemory(r)
```

```
## [1] FALSE
```

```r
nlayers(r)
```

```
## [1] 1
```

```r
dim(r)
```

```
## [1] 115  80   1
```

```r
cellStats(r, function(x, ...) 
  c(min=min(x,...), mean=mean(x,...), max=max(x,...)))
```

```
##       min      mean       max 
##  128.4340  423.1647 1805.7800
```

---

# Operations on `raster` Objects

```r
* r_log <- calc(r, log) 
par(mar=c(5,5,5,5), mfrow=c(1,2))
hist(r); hist(r_log)
```

---

# Operations via External Tools

- `RQGIS`: Uses QGIS Python API and allows you to access many native and 3rd-party plugin based geospatial algorithms/analysis.

- `RQGIS` also allows you to access `GRASS7` and `SAGA`, but you also have option of using `rgrass7` and `RSAGA`.

- `rpostgis` & `postGIStools` allow you to access data in PostGIS, but so does the PostGIS driver in `gdal` so you can query directly from `rgdal` and `sf` too.

---

class: inverse middle

# Part 4: The End!

Continue to [Part 5: Ways to Output Maps](05-Ways_to_Output_Maps.html)

.footnote[Restart [Part 4: Spatial Data Operations](04-Spatial_Data_Operations.html)]