Lecture 04

class: inverse, center, title-slide, middle

# Lecture 04: Spatial Data

## Theory and Tools (a.k.a. GIS Tools Lab.)

### <img src="figs/unibo_logo_white.png" style="width: 15%" /> Bruno Conte

### 28/Sep/2023

---

# Spatial data in economics: schedule

1. ~~Introduction to (spatial) data and programming in `R`~~ [18.Sep.2023]

2. ~~Spatial data basics: vector data + assignment~~ [21.Sep.2023]

3. ~~Basic operations with vector data + assignment~~ [25.Sep.2023]

4. Geometry operations and miscelanea + follow-up + assignment [28.Sep.2023]
 - Follow-up: double-check course's pace and missing concepts
 - Unary geometry operations: simplifying, centroids, buffers, etc
 - Binary geometry operations: clipping, unions, etc.

5. Raster data and operations + assignment [02.Oct.2023]

6. Take-home exam [03.Nov.2023]

---

# Main references for this class

1. Lovelace, R., Nowosad, J. and Muenchow, J., 2019. **Geocomputation with R.** Chapman and Hall/CRC.
 - Chapter 5 (spatial geometry operations)

2. Pebesma, E., 2018. Simple Features for R: Standardized Support for Spatial Vector Data. The R Journal 10 (1), 439-446

3. Wickham, H. and Grolemund, G., 2016. R for data science: import, tidy, transform, visualize, and model data. " O'Reilly Media, Inc.".

---

# Follow-up: any feedback/issues?

- Course pace, contents, (lack of) complexity?

- Timetable, assignments, examination/grading?

- Course format (exposition + practice)?

- The instructor? ;)

- Anything else (email if not comfortable)?

---

# Geometry operations

- Geometry operations: manipulation of vector data that uses/manipulate its geometry. Operations can be both:
 - **Unary:** geometry operations that require (and manipulate) a single feature
 1. Simplification
 2. Centroids
 3. Buffers
 4. Casting
 
 - **Binary:** operations that interact two features (e.g. distance)
 1. Clipping/subsetting
 2. Distances

---

# Unary geometry operations: simplify

.pull-left[
.center[
**Raw geometry:**
![](00_class04_files/figure-html/unnamed-chunk-1-1.png)
]
]
.pull.right[
.center[
**Simplified geometry:**
![](00_class04_files/figure-html/unnamed-chunk-2-1.png)
]
]

It reduces number of vertices. It can be very **CPU intense** (watch out!). Simplifying complex polygons: chapter 5.2.1.

---

# Unary geometry operations: centroids

Assigns the center of mass (**a point**) of a geometry (line or polygon). For disjoint centroids: use `st_point_on_surface()`!

.pull-left[

```r
sf.us <- us_states
*sf.cent <- st_centroid(sf.us)
ggplot() +
 geom_sf(data = sf.us) +
 geom_sf(data = sf.cent, shape = 3)
```
]

.pull-right[
<img src="00_class04_files/figure-html/unnamed-chunk-4-1.png" width="100%" />
]

---

# Unary geometry operations: centroids

Assigns the center of mass (**a point**) of a geometry (line or polygon). For disjoint centroids: use `st_point_on_surface()`!

.pull-left[

```r
sf.river <- seine
*sf.cent <- st_centroid(sf.river)
ggplot() +
 geom_sf(data = sf.river) +
 geom_sf(data = sf.cent,
 shape = 16,
 color = 'red'
 )
```
]

.pull-right[
<img src="00_class04_files/figure-html/unnamed-chunk-6-1.png" width="80%" />
]

---

# Unary geometry operations: centroids

Assigns the center of mass (**a point**) of a geometry (line or polygon). For disjoint centroids: use `st_point_on_surface()`!

.pull-left[

```r
sf.river <- seine
*sf.cent <- st_point_on_surface(sf.river)
ggplot() +
 geom_sf(data = sf.river) +
 geom_sf(data = sf.cent,
 shape = 16,
 color = 'red'
 )
```
]

.pull-right[
<img src="00_class04_files/figure-html/unnamed-chunk-8-1.png" width="80%" />
]

---

# Unary geometry operations: buffers

.pull-left[
- **Buffer zone:** area around a feature containing locations/space within a certain distance.

```r
sf.river <- seine
*sf.buff <- st_buffer(sf.river,
* dist = 5000
* )
ggplot() + 
 geom_sf(data = sf.river) + 
 geom_sf(data = sf.buff, 
 fill = 'red',
 alpha=.5
 )
```

]
.pull.right[
<img src="00_class04_files/figure-html/unnamed-chunk-10-1.png" width="50%" />
]

---

# Unary geometry operations: type transformations (casting)

- **Casting:** transform a feature geometry's type into another based on its vertices

```r
sf.river <- seine
sf.river
```

```
## Simple feature collection with 3 features and 1 field
## Geometry type: MULTILINESTRING
## Dimension:     XY
## Bounding box:  xmin: 518344.7 ymin: 6660431 xmax: 879955.3 ymax: 6938864
## Projected CRS: RGF93 / Lambert-93
##    name                       geometry
## 1 Marne MULTILINESTRING ((879955.3 ...
## 2 Seine MULTILINESTRING ((828893.6 ...
## 3 Yonne MULTILINESTRING ((773482.1 ...
```

---

# Unary geometry operations: type transformations (casting)

- **Casting:** transform a feature geometry's type into another based on its vertices

```r
sf.river <- seine
*sf.cast <- st_cast(sf.river,
* to = 'MULTIPOINT'
* )
sf.cast
```

```
## Simple feature collection with 3 features and 1 field
## Geometry type: MULTIPOINT
## Dimension:     XY
## Bounding box:  xmin: 518344.7 ymin: 6660431 xmax: 879955.3 ymax: 6938864
## Projected CRS: RGF93 / Lambert-93
##    name                       geometry
## 1 Marne MULTIPOINT ((879955.3 67557...
## 2 Seine MULTIPOINT ((828893.6 67138...
## 3 Yonne MULTIPOINT ((773482.1 66604...
```

---

# Binary geometry operations: clipping

**Clipping:** restricting geometry space within the topological relationship between two features. Example - intersection:

.center[<img src="https://r.geocompx.org/05-geometry-operations_files/figure-html/points-1.png" style="width: 50%" />]

---

# Binary geometry operations: clipping

**Clipping:** restricting geometry space within the topological relationship between two features. Example - intersection:

.center[<img src="https://r.geocompx.org/05-geometry-operations_files/figure-html/circle-intersection-1.png" style="width: 50%" />]

---

# Binary geometry operations: clipping (other relationships)

.center[<img src="https://r.geocompx.org/05-geometry-operations_files/figure-html/venn-clip-1.png" style="width: 75%" />]

---

# Binary geometry operations: distances

One of the **most used GIS tools** for economics, with `st_distance()`. Works both with a single or a pair of geometries!

```r
library(spData)
library(sf)
sf.centr <- st_centroid(seine)
*st_distance(sf.centr)
```

```
## Units: [m]
##          1        2        3
## 1      0.0 111458.4 119082.6
## 2 111458.4      0.0 136560.9
## 3 119082.6 136560.9      0.0
```

---

# Binary geometry operations: distances

One of the **most used GIS tools** for economics, with `st_distance()`. Works both with a single or a pair of geometries!

```r
library(spData)
library(sf)
sf.centr <- st_centroid(seine)
*st_distance(sf.centr, by_element = T)
```

```
## Units: [m]
## [1] 0 0 0
```

---

# Binary geometry operations: distances

One of the **most used GIS tools** for economics, with `st_distance()`. Works both with a single or a pair of geometries!

```r
library(spData)
library(sf)
sf.centr <- st_centroid(seine)
*st_distance(sf.centr, seine)
```

```
## Units: [m]
##            [,1]     [,2]      [,3]
## [1,]  32317.409 27135.47 77735.006
## [2,]   3262.824  6126.58 51637.691
## [3,] 129403.368 72121.86  5876.571
```

---

# Binary geometry operations: distances

One of the **most used GIS tools** for economics, with `st_distance()`. Works both with a single or a pair of geometries!

```r
library(spData)
library(sf)
sf.centr <- st_centroid(seine)
*st_distance(sf.centr, seine, by_element = T)
```

```
## Units: [m]
## [1] 32317.409  6126.580  5876.571
```

---
class: inverse, center, middle
count: false

# Your turn: Take-home

# Assignment

---

# Take-home assignment (1/2)

.pull-left[
**Combine** the world shape `world` with:

- Population (point) data (do not use rasters!)

- Ports, airports, etc.

- All data is available at Natural Earth!
]
.pull-right[

**Produce:**

- Map of total population by country

- Histogram of country population distribution by continent

- Histogram of (country-level) average distances between locations and ports **or** airports by continent.

]

---

# Take-home assignment (2/2)

Briefly read

* Porteous, O., 2019. High trade costs and their consequences: An estimated dynamic model of African agricultural storage and trade. *American Economic Journal: Applied Economics*, *11(4)*, pp.327-66.

Then, download the paper's data ([here](https://www.openicpsr.org/openicpsr/project/116359/version/V1/view)), combine with `world` and transportation (e.g. road, railroad) data to:

1. Generate an `sf` `POINT` feature of market prices across Africa; plot them

2. Calculate minimum distance of each market to the (i) coast, (ii) nearest road, and (iii) nearest airport

3. Produce 3 scatter plots relating average prices with the minimum distances

Note: the [paper appendix](https://www.aeaweb.org/doi/10.1257/app.20170442.appx) contains a lot of useful information about the market price data!