Tidy Geospatial Networks in R

# Tidy Geospatial Networks in R
## Introducing the <code>sfnetworks</code> package
### Lucas van der Meer, Lorena Abad, Andrea Gilardi, Robin Lovelace
### 2020-06-16

---

<STYLE type='text/css' scoped>
PRE.fansi SPAN {padding-top: .25em; padding-bottom: .25em};
</STYLE>

### Want to follow along?

Get the slides here: https://sfnetworks.github.io/sfnetworks-webinar/slides
 
--

and/or star this repo: https://github.com/sfnetworks/sfnetworks-webinar
---
class: center, middle
## Hello from the team!
.center[
![:scale 20%](https://avatars1.githubusercontent.com/u/26540305?s=400&u=c576e87314499815cbf698b7781ee58fd1d773e2&v=4)
![:scale 20%](https://pbs.twimg.com/profile_images/1107610946623754240/BaNA8k1E_400x400.png)
![:scale 20%](https://avatars1.githubusercontent.com/u/22221146?s=400&u=3683a04d3f40823162d6c1ae5c51b6c9a5c0a9a5&v=4)
![:scale 20%](https://avatars2.githubusercontent.com/u/10034237?s=460&u=53193bed2fad4f0808b55a227f99897a8d63ebc2&v=4)
]
--
.center[
what brought us together?
]
---
class: center, middle

## Geospatial Networks

--
.pull-left[
Road networks
![](figs/road_network.png)
]

--
.pull-right[
River networks
![](figs/river_network.png)
]

[**spnethack**](https://github.com/sfnetworks/spnethack) organised in Munster, 2019

Building on Luuk's MSc project while Robin was visiting [ifgi](https://www.uni-muenster.de/Geoinformatics/en/)

Result: [Spatial networks in R with sf and tidygraph](https://www.r-spatial.org/r/2019/09/26/spatial-networks.html) blog post

---

- [\#rspatial](https://twitter.com/search?q=%23rspatial)
- `sf`
- `stars`
- `rgeos`
- `rgdal`
- `tmap`
- `ggmap`
- `mapview`
- ...

]

]
---

- [statnet](http://statnet.org/)
- `igraph`
- `tidygraph`
- `qgraph`
- `ggraph`
- `visNetwork`
- `networkD3`
- ...
]

---
class: center, middle

# then...

## why a new package?

---

[![](figs/so_q1.png)](https://stackoverflow.com/questions/57833905/how-to-convert-a-list-of-sf-spatial-points-into-a-routable-graph)

]

---

[![:scale 80%](figs/gh_q1.png)](https://github.com/r-spatial/sf/issues/966)
]

---

[![](figs/gh_c1.png)](https://github.com/r-spatial/sf/issues/790#issuecomment-403831517)
]

---
class: center
## why a new package?

<center><blockquote class="twitter-tweet">One of the biggest reasons we still have ArcGIS licenses is for Network Analysis (drive times, service areas etc). Does anyone have <a href="https://twitter.com/hashtag/foss4g?src=hash&amp;ref_src=twsrc%5Etfw">#foss4g</a> tools for this they like? Last time we tried pgRouting (yrs ago) it didn&#39;t feel fully formed yet, <a href="https://twitter.com/hashtag/gischat?src=hash&amp;ref_src=twsrc%5Etfw">#gischat</a>&mdash; Zev Ross (@zevross) <a href="https://twitter.com/zevross/status/1089908839816794118?ref_src=twsrc%5Etfw">January 28, 2019</a></blockquote></center> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script>

...

because open source can always get better!

---
class: center, middle

## What is out there already?

[dodgr](https://atfutures.github.io/dodgr/)

[cppRouting](https://github.com/vlarmet/cppRouting)

[shp2graph](https://r-forge.r-project.org/projects/shp2graph)

[spnetwork](https://github.com/edzer/spnetwork)

[stplanr](https://docs.ropensci.org/stplanr/)

...

---
class: center, middle

## What do we propose?

--
.pull-left[
![](https://user-images.githubusercontent.com/520851/34887433-ce1d130e-f7c6-11e7-83fc-d60ad4fae6bd.gif)
]

.pull-right[
![:scale 38%](https://raw.githubusercontent.com/thomasp85/tidygraph/master/man/figures/logo.png)
]
--

## `sfnetworks`

The best of both worlds!
---
class: middle

.pull-left-70[
.center[
![:scale 80%](https://github.com/allisonhorst/stats-illustrations/raw/master/rstats-artwork/tidyverse_celestial.png)
]
.footnote[
Artwork by [@allison_horst](https://twitter.com/allison_horst)
]
]

.pull-right-30[
### *tidy* workflows
> In tidy data:
1. Each variable forms a column.
2. Each observation forms a row.
3. Each type of observational unit forms a table.

.footnote[
Wickham, H. (2014). Tidy Data. Journal of Statistical Software, 59(10), 1 - 23. doi:http://dx.doi.org/10.18637/jss.v059.i10
]

- Supported by the *tidyverse*

- Allows piping `%>%` structures

]

---
class: center, middle

.pull-left-70[
![](https://github.com/allisonhorst/stats-illustrations/raw/master/rstats-artwork/sf.png)
.footnote[
Artwork by [@allison_horst](https://twitter.com/allison_horst)
]
]

### **`sf` package**

Simple features for R

Spatial vector data (points, lines and polygons)

Compatible with *tidy* workflows

S3 classes
]
---
class: center, middle

.pull-left-70[
![:scale 80%](https://www.r-graph-gallery.com/339-circular-dendrogram-with-ggraph_files/figure-html/thecode8-1.png)
.footnote[
Plot from [R GRaph Gallery](https://www.r-graph-gallery.com/339-circular-dendrogram-with-ggraph.html)
]
]

### **`tidygraph` package**
![:scale 10%](https://avatars3.githubusercontent.com/u/3735184?s=200&v=4)
 Interfaces with [`igraph`](https://igraph.org/r/)  
 
![:scale 8%](https://github.com/tidyverse/dplyr/raw/master/man/figures/logo.png)
 Supports [`dplyr`](https://dplyr.tidyverse.org/) *verbs*

![:scale 8%](https://raw.githubusercontent.com/thomasp85/tidygraph/master/man/figures/logo.png)
 Introduces new *verbs* specific to network data (e.g. `morph`, `bind_graphs`, `graph_join`)

![:scale 8%](https://github.com/thomasp85/ggraph/raw/master/man/figures/logo.png)
 Allows network visualization via [`ggraph`](https://ggraph.data-imaginist.com/)
]

---
class: center, middle

## So, let's dive in...

---
### Installation

Install the more stable master branch with:

```r
remotes::install_github("luukvdmeer/sfnetworks")
```

Install the develop branch, where most of the development takes place and where PRs should be directed:

```r
remotes::install_github("luukvdmeer/sfnetworks", ref = "develop")
```

---
## `sfnetwork` data structure

### Philosophy

> "... a close approximation of tidyness for relational data is two tidy data frames, one describing the node data and one describing the edge data."

`tidygraph` introduction.

> “A close approximation of tidyness for relational .orange[*geospatial data*] is two .orange[*sf objects*], one describing the node data and one describing the edge data.”

`sfnetworks` adaption.

---
## `sfnetwork` data structure

### Construction

- Nodes: `sf` object with `POINT` geometries

- Edges: *to* and *from* column with end-points

- Same CRS!

---
class: split-40
count: false

```r
*p1 = st_point(c(7, 51))
*p2 = st_point(c(7, 52))
*p3 = st_point(c(8, 52))

*st_sfc(
*   p1, p2, p3,
*   crs = 4326
* ) %>% st_sf()
```
]
 
.right-panel-toyexample-user[

```
Simple feature collection with 3 features and 0 fields
geometry type:  POINT
dimension:      XY
bbox:           xmin: 7 ymin: 51 xmax: 8 ymax: 52
geographic CRS: WGS 84
      geometry
1 POINT (7 51)
2 POINT (7 52)
3 POINT (8 52)
```
]

---
class: split-40
count: false

```r
p1 = st_point(c(7, 51))  
p2 = st_point(c(7, 52))  
p3 = st_point(c(8, 52))

st_sfc(  
    p1, p2, p3,  
    crs = 4326  
  ) %>% st_sf() ->  
* nodes
```
]
 
.right-panel-toyexample-user[

]

---
class: split-40
count: false

```r
p1 = st_point(c(7, 51))  
p2 = st_point(c(7, 52))  
p3 = st_point(c(8, 52))

st_sfc(  
    p1, p2, p3,  
    crs = 4326  
  ) %>% st_sf() ->  
  nodes

*st_sfc(
* st_cast(st_union(p1,p2), "LINESTRING"),
* st_cast(st_union(p1,p3), "LINESTRING"),
* st_cast(st_union(p2,p3), "LINESTRING"),
* crs = 4326
*) %>% st_sf()
```
]
 
.right-panel-toyexample-user[

```
Simple feature collection with 3 features and 0 fields
geometry type:  LINESTRING
dimension:      XY
bbox:           xmin: 7 ymin: 51 xmax: 8 ymax: 52
geographic CRS: WGS 84
                 geometry
1 LINESTRING (7 51, 7 52)
2 LINESTRING (7 51, 8 52)
3 LINESTRING (7 52, 8 52)
```
]

---
class: split-40
count: false

```r
p1 = st_point(c(7, 51))  
p2 = st_point(c(7, 52))  
p3 = st_point(c(8, 52))

st_sfc(  
    p1, p2, p3,  
    crs = 4326  
  ) %>% st_sf() ->  
  nodes

st_sfc(  
  st_cast(st_union(p1,p2), "LINESTRING"),  
  st_cast(st_union(p1,p3), "LINESTRING"),  
  st_cast(st_union(p2,p3), "LINESTRING"),  
  crs = 4326  
) %>% st_sf() ->  
* edges
```
]
 
.right-panel-toyexample-user[

]

---
class: split-40
count: false

```r
p1 = st_point(c(7, 51))  
p2 = st_point(c(7, 52))  
p3 = st_point(c(8, 52))

st_sfc(  
    p1, p2, p3,  
    crs = 4326  
  ) %>% st_sf() ->  
  nodes

st_sfc(  
  st_cast(st_union(p1,p2), "LINESTRING"),  
  st_cast(st_union(p1,p3), "LINESTRING"),  
  st_cast(st_union(p2,p3), "LINESTRING"),  
  crs = 4326  
) %>% st_sf() ->  
  edges

*edges$from = c(1, 1, 2)
*edges$to = c(2, 3, 3)
```
]
 
.right-panel-toyexample-user[

]

---
class: split-40
count: false

```r
p1 = st_point(c(7, 51))  
p2 = st_point(c(7, 52))  
p3 = st_point(c(8, 52))

st_sfc(  
    p1, p2, p3,  
    crs = 4326  
  ) %>% st_sf() ->  
  nodes

st_sfc(  
  st_cast(st_union(p1,p2), "LINESTRING"),  
  st_cast(st_union(p1,p3), "LINESTRING"),  
  st_cast(st_union(p2,p3), "LINESTRING"),  
  crs = 4326  
) %>% st_sf() ->  
  edges

edges$from = c(1, 1, 2)  
edges$to = c(2, 3, 3)

*sfnetwork(nodes, edges, directed = FALSE)
```
]
 
.right-panel-toyexample-user[

```
Checking validity of network structure... Use force=TRUE to force construction without checks
```

<PRE class="fansi fansi-output"><CODE># An sfnetwork with 3 nodes and 3 edges
#
# CRS: EPSG:4326 
#
# An undirected simple graph with 1 component with spatially explicit edges
#
# Node Data: 3 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7 ymin: 51 xmax: 8 ymax: 52
 geometry
 &lt;POINT [°]&gt;
1 (7 51)
2 (7 52)
3 (8 52)
#
# Edge Data: 3 x 3
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7 ymin: 51 xmax: 8 ymax: 52
 from to geometry
 &lt;int&gt; &lt;int&gt; &lt;LINESTRING [°]&gt;
1 1 2 (7 51, 7 52)
2 1 3 (7 51, 8 52)
3 2 3 (7 52, 8 52)
</CODE></PRE>
]

---
## `sfnetwork` data structure

### Foreign objects

To convert foreign objects into an `sfnetwork` with `as_sfnetwork`, there are two conditions:

- Object is or can be converted to a `tbl_graph`

- At least the nodes can be converted to an `sf` object

Or just an `sf` object with `POINT` or `LINESTRING` geometry

---

```r
*roxel
```
]
 
.right-panel-object1-user[

```
Simple feature collection with 851 features and 2 fields
geometry type: LINESTRING
dimension: XY
bbox: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax: 51.9612
geographic CRS: WGS 84
# A tibble: 851 x 3
 name type geometry
 <fct> <fct> <LINESTRING [°]>
 1 Havixbecker ~ residen~ (7.533722 51.95556, 7.533461 51.95576)
 2 Pienersallee seconda~ (7.532442 51.95422, 7.53236 51.95377, 7.532~
 3 Schulte-Bern~ residen~ (7.532709 51.95209, 7.532823 51.95239, 7.53~
 4 <NA> path (7.540063 51.94468, 7.539696 51.94479, 7.53~
 5 Welsingheide residen~ (7.537673 51.9475, 7.537614 51.94562)
 6 <NA> footway (7.543791 51.94733, 7.54369 51.94686, 7.543~
 7 <NA> footway (7.54012 51.94478, 7.539931 51.94514)
 8 <NA> path (7.53822 51.94546, 7.538131 51.94549, 7.538~
 9 <NA> track (7.540063 51.94468, 7.540338 51.94468, 7.54~
10 <NA> track (7.5424 51.94599, 7.54205 51.94629, 7.54196~
# ... with 841 more rows
```
]

---
class: split-40
count: false

```r
roxel %>%  
* as_sfnetwork()
```
]
 
.right-panel-object1-user[

```
# An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# A directed multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 <POINT [°]>
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 <int> <int> <fct> <fct> <LINESTRING [°]>
1 1 2 Havixbecke~ reside~ (7.533722 51.95556, 7.533461 51.955~
2 3 4 Pienersall~ second~ (7.532442 51.95422, 7.53236 51.9537~
3 5 6 Schulte-Be~ reside~ (7.532709 51.95209, 7.532823 51.952~
# ... with 848 more rows
```
]

---
class: split-40
count: false

```r
roxel %>%  
  as_sfnetwork() ->  
* net
```
]
 
.right-panel-object1-user[

]

---

```r
*net %>%
* st_crs()
```
]
 
.right-panel-object2-user[

```
Coordinate Reference System:
  User input: EPSG:4326 
  wkt:
GEOGCRS["WGS 84",
    DATUM["World Geodetic System 1984",
        ELLIPSOID["WGS 84",6378137,298.257223563,
            LENGTHUNIT["metre",1]]],
    PRIMEM["Greenwich",0,
        ANGLEUNIT["degree",0.0174532925199433]],
    CS[ellipsoidal,2],
        AXIS["geodetic latitude (Lat)",north,
            ORDER[1],
            ANGLEUNIT["degree",0.0174532925199433]],
        AXIS["geodetic longitude (Lon)",east,
            ORDER[2],
            ANGLEUNIT["degree",0.0174532925199433]],
    USAGE[
        SCOPE["unknown"],
        AREA["World"],
        BBOX[-90,-180,90,180]],
    ID["EPSG",4326]]
```
]

---
class: split-40
count: false

```r
net %>%  
  st_crs()

*net %>%
* st_transform(3035)
```
]
 
.right-panel-object2-user[

```
# An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:3035 
#
# A directed multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 geometry
 <POINT [m]>
1 (4151491 3207923)
2 (4151474 3207946)
3 (4151398 3207777)
4 (4151370 3207673)
5 (4151408 3207539)
6 (4151421 3207592)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 from to name type geometry
 <int> <int> <fct> <fct> <LINESTRING [m]>
1 1 2 Havixbecker~ reside~ (4151491 3207923, 4151474 3207946)
2 3 4 Pienersallee second~ (4151398 3207777, 4151390 3207727,~
3 5 6 Schulte-Ber~ reside~ (4151408 3207539, 4151417 3207573,~
# ... with 848 more rows
```
]

---

```r
*net %>%
* sf_attr("sf_column", "nodes")
```
]
 
.right-panel-object4-user[

```
[1] "geometry"
```
]

---
class: split-40
count: false

```r
net %>%  
  sf_attr("sf_column", "nodes")

*net %>%
* sf_attr("agr", "edges")  # BREAK
```
]
 
.right-panel-object4-user[

```
[1] "geometry"
```

```
from to name type 
<NA> <NA> <NA> <NA> 
Levels: constant aggregate identity
```
]

---

```r
*par(mar = c(1, 1, 1, 1), bg = NA)
*plot(net)
```
]
 
.right-panel-object5-user[
<img src="slides_files/figure-html/object5_user_1_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
par(mar = c(1, 1, 1, 1), bg = NA)  
plot(net)

*net %>%
* class()
```
]
 
.right-panel-object5-user[
<img src="slides_files/figure-html/object5_user_2_output-1.png" width="432" />

```
[1] "sfnetwork" "tbl_graph" "igraph"   
```
]

---

`sfnetwork` objects subclass `tbl_graph` objects, so any `igraph` algorithm and tidy wrappers from `tidygraph` are supported.

Check out the possibilities on the [tidygraph introduction](https://www.data-imaginist.com/2017/introducing-tidygraph/).
]

]
---

```r
roxel %>% 
 as_sfnetwork( 
 ) 
```
]
 
.right-panel-sfnetworks_options-non_seq[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# A directed multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havixbecke~ reside~ (7.533722 51.95556, 7.533461 51.955~
2 3 4 Pienersall~ second~ (7.532442 51.95422, 7.53236 51.9537~
3 5 6 Schulte-Be~ reside~ (7.532709 51.95209, 7.532823 51.952~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
roxel %>% 
 as_sfnetwork( 
* directed = F,
 ) 
```
]
 
.right-panel-sfnetworks_options-non_seq[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havixbecke~ reside~ (7.533722 51.95556, 7.533461 51.955~
2 3 4 Pienersall~ second~ (7.532442 51.95422, 7.53236 51.9537~
3 5 6 Schulte-Be~ reside~ (7.532709 51.95209, 7.532823 51.952~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
roxel %>% 
 as_sfnetwork( 
 directed = F, 
* edges_as_lines = F
 ) 
```
]
 
.right-panel-sfnetworks_options-non_seq[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially implicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 4
 from to name type 
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; 
1 1 2 Havixbecker Strasse residential
2 3 4 Pienersallee secondary 
3 5 6 Schulte-Bernd-Strasse residential
# ... with 848 more rows
</CODE></PRE>
]

.note[See issue [#50](https://github.com/luukvdmeer/sfnetworks/issues/50)]
---
class: split-40
count: false

```r
*net
```
]
 
.right-panel-activate-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# A directed multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havixbecke~ reside~ (7.533722 51.95556, 7.533461 51.955~
2 3 4 Pienersall~ second~ (7.532442 51.95422, 7.53236 51.9537~
3 5 6 Schulte-Be~ reside~ (7.532709 51.95209, 7.532823 51.952~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
* activate("edges")
```
]
 
.right-panel-activate-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# A directed multigraph with 14 components with spatially explicit edges
#
# Edge Data: 851 x 5 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havixbeck~ reside~ (7.533722 51.95556, 7.533461 51.9557~
2 3 4 Pienersal~ second~ (7.532442 51.95422, 7.53236 51.95377~
3 5 6 Schulte-B~ reside~ (7.532709 51.95209, 7.532823 51.9523~
4 7 8 NA path (7.540063 51.94468, 7.539696 51.9447~
5 9 10 Welsinghe~ reside~ (7.537673 51.9475, 7.537614 51.94562)
6 11 12 NA footway (7.543791 51.94733, 7.54369 51.94686~
# ... with 845 more rows
#
# Node Data: 701 x 1
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
# ... with 698 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>%  
  activate("edges") %>%  
* st_geometry()
```
]
 
.right-panel-activate-user[

```
Geometry set for 851 features 
geometry type:  LINESTRING
dimension:      XY
bbox:           xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax: 51.9612
geographic CRS: WGS 84
First 5 geometries:
```

```
LINESTRING (7.533722 51.95556, 7.533461 51.95576)
```

```
LINESTRING (7.532442 51.95422, 7.53236 51.95377...
```

```
LINESTRING (7.532709 51.95209, 7.532823 51.9523...
```

```
LINESTRING (7.540063 51.94468, 7.539696 51.9447...
```

```
LINESTRING (7.537673 51.9475, 7.537614 51.94562)
```
]

---
class: split-40
count: false

```r
net %>%  
  activate("edges") %>%  
  st_geometry()

*net %>%
* activate("nodes") %>%
* st_geometry()
```
]
 
.right-panel-activate-user[

```
LINESTRING (7.533722 51.95556, 7.533461 51.95576)
```

```
LINESTRING (7.532442 51.95422, 7.53236 51.95377...
```

```
LINESTRING (7.532709 51.95209, 7.532823 51.9523...
```

```
LINESTRING (7.540063 51.94468, 7.539696 51.9447...
```

```
LINESTRING (7.537673 51.9475, 7.537614 51.94562)
```

```
Geometry set for 701 features 
geometry type:  POINT
dimension:      XY
bbox:           xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax: 51.9612
geographic CRS: WGS 84
First 5 geometries:
```

```
POINT (7.533722 51.95556)
```

```
POINT (7.533461 51.95576)
```

```
POINT (7.532442 51.95422)
```

```
POINT (7.53209 51.95328)
```

```
POINT (7.532709 51.95209)
```
]

---
class: split-40
count: false

```r
*net = roxel %>% as_sfnetwork(directed = F)
```
]
 
.right-panel-filter-user[

]

---
class: split-40
count: false

```r
net = roxel %>% as_sfnetwork(directed = F)

*par(mar = c(1, 1, 1, 1), bg = NA)
*plot(net)
```
]
 
.right-panel-filter-user[
<img src="slides_files/figure-html/filter_user_2_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
net = roxel %>% as_sfnetwork(directed = F)

par(mar = c(1, 1, 1, 1), bg = NA)  
plot(net)

*rect = st_multipoint(c(
*   st_point(c(7.53173, 51.95662)),
*   st_point(c(7.53173, 51.95190)),
*   st_point(c(7.53778, 51.95190)),
*   st_point(c(7.53778, 51.95662))
* )) %>%
* st_cast('POLYGON') %>%
* st_sfc(crs = 4326)

*rect %>%
* plot(
* border = "Red",
* lwd = 2,
* add = TRUE
*) 
```
]
 
.right-panel-filter-user[
<img src="slides_files/figure-html/filter_user_3_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
net = roxel %>% as_sfnetwork(directed = F)

par(mar = c(1, 1, 1, 1), bg = NA)  
plot(net)

rect = st_multipoint(c(  
    st_point(c(7.53173, 51.95662)),  
    st_point(c(7.53173, 51.95190)),  
    st_point(c(7.53778, 51.95190)),  
    st_point(c(7.53778, 51.95662))  
  )) %>%  
  st_cast('POLYGON') %>%  
  st_sfc(crs = 4326)

rect %>%  
  plot(  
  border = "Red",  
  lwd = 2,  
  add = TRUE  
)

*net %>%
* st_filter(rect, .pred = st_intersects) %>%
* plot(col = "Orange", add = TRUE)
```
]
 
.right-panel-filter-user[

```
although coordinates are longitude/latitude, st_intersects assumes that they are planar
although coordinates are longitude/latitude, st_intersects assumes that they are planar
```

<img src="slides_files/figure-html/filter_user_4_output-1.png" width="432" />
]

### Not supported functions

> There is a clear limitation in the relational data structure that requires rows to maintain their identity.

> For sfnetwork objects, this also means that the geometries of features should be maintained, or at least the endpoints of these geometries, in the case of the edges.

> That is, functions that summarise geometries, or (may) change their .orange[type, shape or position], are not supported.

---
class: split-40
count: false

```r
*net %>%
* st_transform(3035) %>%
* activate("nodes")
```
]
 
.right-panel-extraction-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:3035 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 geometry
 &lt;POINT [m]&gt;
1 (4151491 3207923)
2 (4151474 3207946)
3 (4151398 3207777)
4 (4151370 3207673)
5 (4151408 3207539)
6 (4151421 3207592)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [m]&gt;
1 1 2 Havixbecker~ reside~ (4151491 3207923, 4151474 3207946)
2 3 4 Pienersallee second~ (4151398 3207777, 4151390 3207727,~
3 5 6 Schulte-Ber~ reside~ (4151408 3207539, 4151417 3207573,~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
 st_transform(3035) %>% 
 activate("nodes") %>% 
* st_as_sf()
```
]
 
.right-panel-extraction-user[
<PRE class="fansi fansi-output"><CODE>Simple feature collection with 701 features and 0 fields
geometry type: POINT
dimension: XY
bbox: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax: 3208565
projected CRS: ETRS89-extended / LAEA Europe
# A tibble: 701 x 1
 geometry
 &lt;POINT [m]&gt;
 1 (4151491 3207923)
 2 (4151474 3207946)
 3 (4151398 3207777)
 4 (4151370 3207673)
 5 (4151408 3207539)
 6 (4151421 3207592)
 7 (4151885 3206698)
 8 (4151762 3206789)
 9 (4151732 3207017)
10 (4151721 3206809)
# ... with 691 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
 st_transform(3035) %>% 
 activate("nodes") %>% 
 st_as_sf() %>% 
* mutate(
* aoi = st_area(
* st_collection_extract(
* st_voronoi(do.call(c, st_geometry(.)))
* )))
```
]
 
.right-panel-extraction-user[
<PRE class="fansi fansi-output"><CODE>Simple feature collection with 701 features and 1 field
geometry type: POINT
dimension: XY
bbox: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax: 3208565
projected CRS: ETRS89-extended / LAEA Europe
# A tibble: 701 x 2
 geometry aoi
 * &lt;POINT [m]&gt; &lt;dbl&gt;
 1 (4151491 3207923) 1837540.
 2 (4151474 3207946) 2563698.
 3 (4151398 3207777) 36462.
 4 (4151370 3207673) 359624.
 5 (4151408 3207539) 12954.
 6 (4151421 3207592) 11089.
 7 (4151885 3206698) 190473.
 8 (4151762 3206789) 26370.
 9 (4151732 3207017) 8416.
10 (4151721 3206809) 4867.
# ... with 691 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>%  
  st_transform(3035) %>%  
  activate("nodes") %>%  
  st_as_sf() %>%  
  mutate(  
    aoi = st_area(  
      st_collection_extract(  
        st_voronoi(do.call(c, st_geometry(.)))  
  ))) ->  
* y

*net %>%
* st_transform(3035) %>%
* activate("nodes")
```
]
 
.right-panel-extraction-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:3035 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 geometry
 &lt;POINT [m]&gt;
1 (4151491 3207923)
2 (4151474 3207946)
3 (4151398 3207777)
4 (4151370 3207673)
5 (4151408 3207539)
6 (4151421 3207592)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [m]&gt;
1 1 2 Havixbecker~ reside~ (4151491 3207923, 4151474 3207946)
2 3 4 Pienersallee second~ (4151398 3207777, 4151390 3207727,~
3 5 6 Schulte-Ber~ reside~ (4151408 3207539, 4151417 3207573,~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>%  
  st_transform(3035) %>%  
  activate("nodes") %>%  
  st_as_sf() %>%  
  mutate(  
    aoi = st_area(  
      st_collection_extract(  
        st_voronoi(do.call(c, st_geometry(.)))  
  ))) ->  
  y

net %>% 
 st_transform(3035) %>% 
 activate("nodes") %>% 
* st_join(y)
```
]
 
.right-panel-extraction-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:3035 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 2 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 geometry aoi
 &lt;POINT [m]&gt; &lt;dbl&gt;
1 (4151491 3207923) 1837540.
2 (4151474 3207946) 2563698.
3 (4151398 3207777) 36462.
4 (4151370 3207673) 359624.
5 (4151408 3207539) 12954.
6 (4151421 3207592) 11089.
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [m]&gt;
1 1 2 Havixbecker~ reside~ (4151491 3207923, 4151474 3207946)
2 3 4 Pienersallee second~ (4151398 3207777, 4151390 3207727,~
3 5 6 Schulte-Ber~ reside~ (4151408 3207539, 4151417 3207573,~
# ... with 848 more rows
</CODE></PRE>
]

---
class: center, middle

## Extending the vocabulary

---
class: split-40
count: false

```r
net %>% 
 activate("edges") %>% 
 mutate(length = edge_length()) 
```
]
 
.right-panel-edgemeasures-rotate[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Edge Data: 851 x 6 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry length
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt; [m]
1 1 2 Havixbe~ resid~ (7.533722 51.95556, 7.533461 5~ 28.859~
2 3 4 Pieners~ secon~ (7.532442 51.95422, 7.53236 51~ 107.704~
3 5 6 Schulte~ resid~ (7.532709 51.95209, 7.532823 5~ 54.362~
4 7 8 NA path (7.540063 51.94468, 7.539696 5~ 155.230~
5 9 10 Welsing~ resid~ (7.537673 51.9475, 7.537614 51~ 208.663~
6 11 12 NA footw~ (7.543791 51.94733, 7.54369 51~ 63.023~
# ... with 845 more rows
#
# Node Data: 701 x 1
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
# ... with 698 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
 activate("edges") %>% 
* mutate(straight = edge_straight_length())
```
]
 
.right-panel-edgemeasures-rotate[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Edge Data: 851 x 6 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry straight
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt; [m]
1 1 2 Havixbe~ resid~ (7.533722 51.95556, 7.533461 5~ 28.859~
2 3 4 Pieners~ secon~ (7.532442 51.95422, 7.53236 51~ 107.073~
3 5 6 Schulte~ resid~ (7.532709 51.95209, 7.532823 5~ 54.336~
4 7 8 NA path (7.540063 51.94468, 7.539696 5~ 153.916~
5 9 10 Welsing~ resid~ (7.537673 51.9475, 7.537614 51~ 208.663~
6 11 12 NA footw~ (7.543791 51.94733, 7.54369 51~ 61.702~
# ... with 845 more rows
#
# Node Data: 701 x 1
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
# ... with 698 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
 activate("edges") %>% 
* mutate(circuity = edge_circuity())
```
]
 
.right-panel-edgemeasures-rotate[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Edge Data: 851 x 6 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry circuity
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt; [1]
1 1 2 Havixbe~ resid~ (7.533722 51.95556, 7.533461 5~ 1.000000
2 3 4 Pieners~ secon~ (7.532442 51.95422, 7.53236 51~ 1.005896
3 5 6 Schulte~ resid~ (7.532709 51.95209, 7.532823 5~ 1.000472
4 7 8 NA path (7.540063 51.94468, 7.539696 5~ 1.008539
5 9 10 Welsing~ resid~ (7.537673 51.9475, 7.537614 51~ 1.000000
6 11 12 NA footw~ (7.543791 51.94733, 7.54369 51~ 1.021405
# ... with 845 more rows
#
# Node Data: 701 x 1
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
# ... with 698 more rows
</CODE></PRE>
]

---

```r
*net_proj = net %>% st_transform(3035)

*p1 = net_proj %>%
* activate(nodes) %>%
* st_as_sf() %>%
* slice(1)

*p2 = net_proj %>%
* activate(nodes) %>%
* st_as_sf() %>%
* slice(9)
```
]
 
.right-panel-shortestpath-user[

]

---
class: split-40
count: false

```r
net_proj = net %>% st_transform(3035)

p1 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(1)

p2 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(9)

*p3 = sf::st_sfc(
* sf::st_geometry(p1)[[1]] + sf::st_point(c(500, 500)),
* crs = sf::st_crs(p1)
*)

*p4 = sf::st_sfc(
* sf::st_geometry(p2)[[1]] + sf::st_point(c(-500, -500)),
* crs = sf::st_crs(p2)
*) 
```
]
 
.right-panel-shortestpath-user[

]

---
class: split-40
count: false

```r
net_proj = net %>% st_transform(3035)

p1 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(1)

p2 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(9)

p3 = sf::st_sfc(  
  sf::st_geometry(p1)[[1]] + sf::st_point(c(500, 500)),  
  crs = sf::st_crs(p1)  
)

p4 = sf::st_sfc(  
  sf::st_geometry(p2)[[1]] + sf::st_point(c(-500, -500)),  
  crs = sf::st_crs(p2)  
)

*par(mar = c(1,1,1,1), bg = NA)
*plot(net_proj)
```
]
 
.right-panel-shortestpath-user[
<img src="slides_files/figure-html/shortestpath_user_3_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
net_proj = net %>% st_transform(3035)

p1 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(1)

p2 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(9)

p3 = sf::st_sfc(  
  sf::st_geometry(p1)[[1]] + sf::st_point(c(500, 500)),  
  crs = sf::st_crs(p1)  
)

p4 = sf::st_sfc(  
  sf::st_geometry(p2)[[1]] + sf::st_point(c(-500, -500)),  
  crs = sf::st_crs(p2)  
)

par(mar = c(1,1,1,1), bg = NA) 
plot(net_proj) 
*plot(p3, col = "red", pch = 20, cex = 2, add = TRUE)
```
]
 
.right-panel-shortestpath-user[
<img src="slides_files/figure-html/shortestpath_user_4_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
net_proj = net %>% st_transform(3035)

p1 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(1)

p2 = net_proj %>%  
  activate(nodes) %>%  
  st_as_sf() %>%  
  slice(9)

p3 = sf::st_sfc(  
  sf::st_geometry(p1)[[1]] + sf::st_point(c(500, 500)),  
  crs = sf::st_crs(p1)  
)

p4 = sf::st_sfc(  
  sf::st_geometry(p2)[[1]] + sf::st_point(c(-500, -500)),  
  crs = sf::st_crs(p2)  
)

par(mar = c(1,1,1,1), bg = NA) 
plot(net_proj) 
plot(p3, col = "red", pch = 20, cex = 2, add = TRUE) 
*plot(p4, col = "red", pch = 20, cex = 2, add = TRUE)
```
]
 
.right-panel-shortestpath-user[
<img src="slides_files/figure-html/shortestpath_user_5_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
*net_proj %>%
* activate("edges") %>%
* mutate(weight = edge_length())
```
]
 
.right-panel-shortestpath1-user[

```
# An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:3035 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Edge Data: 851 x 6 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 from to name type geometry weight
 <int> <int> <fct> <fct> <LINESTRING [m]> [m]
1 1 2 Havixbe~ resid~ (4151491 3207923, 4151474 3207~ 28.859~
2 3 4 Pieners~ secon~ (4151398 3207777, 4151390 3207~ 107.714~
3 5 6 Schulte~ resid~ (4151408 3207539, 4151417 3207~ 54.367~
4 7 8 <NA> path (4151885 3206698, 4151861 3206~ 155.223~
5 9 10 Welsing~ resid~ (4151732 3207017, 4151721 3206~ 208.682~
6 11 12 <NA> footw~ (4152152 3206984, 4152143 3206~ 63.028~
# ... with 845 more rows
#
# Node Data: 701 x 1
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150707 ymin: 3206375 xmax: 4152367 ymax:
# 3208565
 geometry
 <POINT [m]>
1 (4151491 3207923)
2 (4151474 3207946)
3 (4151398 3207777)
# ... with 698 more rows
```
]

---
class: split-40
count: false

```r
net_proj %>%  
  activate("edges") %>%  
  mutate(weight = edge_length()) %>%  
* st_shortest_paths(p3, p4) %>%
* .$vpath
```
]
 
.right-panel-shortestpath1-user[

```
[[1]]
+ 45/701 vertices, from 90c0122:
 [1] 178 179 601 121  29  28 478 140 135 136 143 471 579 263 262 480
[17]   3   4 248 113  63 321 210 470 319 317 323 550 554 546  67 194
[33] 646 439 440 565 666 701 161 628 242 493 240 157 158
```
]

---

```r
*net %>%
* morph(to_components) ->
* net_m
```
]
 
.right-panel-morphers-user[

]

---
class: split-40
count: false

```r
net %>%  
  morph(to_components) ->  
  net_m

*net_m %>% class()
```
]
 
.right-panel-morphers-user[

```
[1] "morphed_sfnetwork" "morphed_tbl_graph" "list"             
```
]

---
class: split-40
count: false

```r
net %>%  
  morph(to_components) ->  
  net_m

net_m %>% class()

*net_m %>% length()
```
]
 
.right-panel-morphers-user[

```
[1] "morphed_sfnetwork" "morphed_tbl_graph" "list"             
```

```
[1] 14
```
]

---
class: split-40
count: false

```r
net %>%  
  morph(to_components) ->  
  net_m

net_m %>% class()

net_m %>% length()

*net %>%
* convert(to_components, .select = 1)
```
]
 
.right-panel-morphers-user[

```
[1] "morphed_sfnetwork" "morphed_tbl_graph" "list"             
```

```
[1] 14
```

<PRE class="fansi fansi-output"><CODE># An sfnetwork with 675 nodes and 838 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 1 component with spatially explicit edges
#
# Node Data: 675 x 2 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry .tidygraph_node_index
 &lt;POINT [°]&gt; &lt;int&gt;
1 (7.533722 51.95556) 1
2 (7.533461 51.95576) 2
3 (7.532442 51.95422) 3
4 (7.53209 51.95328) 4
5 (7.532709 51.95209) 5
6 (7.532869 51.95257) 6
# ... with 669 more rows
#
# Edge Data: 838 x 6
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry .tidygraph_edge~
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt; &lt;int&gt;
1 1 2 Havix~ resi~ (7.533722 51.95556, 7.533~ 1
2 3 4 Piene~ seco~ (7.532442 51.95422, 7.532~ 2
3 5 6 Schul~ resi~ (7.532709 51.95209, 7.532~ 3
# ... with 835 more rows
</CODE></PRE>
]

---

```r
*net
```
]
 
.right-panel-spatialmorphers1-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 701 nodes and 851 edges
#
# CRS: EPSG:4326 
#
# An undirected multigraph with 14 components with spatially explicit edges
#
# Node Data: 701 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522622 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.532442 51.95422)
4 (7.53209 51.95328)
5 (7.532709 51.95209)
6 (7.532869 51.95257)
# ... with 695 more rows
#
# Edge Data: 851 x 5
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havixbecke~ reside~ (7.533722 51.95556, 7.533461 51.955~
2 3 4 Pienersall~ second~ (7.532442 51.95422, 7.53236 51.9537~
3 5 6 Schulte-Be~ reside~ (7.532709 51.95209, 7.532823 51.952~
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
* convert(to_spatial_coordinates)
```
]
 
.right-panel-spatialmorphers1-user[
<PRE class="fansi fansi-output"><CODE># A tbl_graph: 701 nodes and 851 edges
#
# An undirected multigraph with 14 components
#
# Node Data: 701 x 3 (active)
 .tidygraph_node_index X Y
 &lt;int&gt; &lt;dbl&gt; &lt;dbl&gt;
1 1 7.53 52.0
2 2 7.53 52.0
3 3 7.53 52.0
4 4 7.53 52.0
5 5 7.53 52.0
6 6 7.53 52.0
# ... with 695 more rows
#
# Edge Data: 851 x 5
 from to name type .tidygraph_edge_index
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;int&gt;
1 1 2 Havixbecker Strasse residential 1
2 3 4 Pienersallee secondary 2
3 5 6 Schulte-Bernd-Strasse residential 3
# ... with 848 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
*net %>%
* convert(to_spatial_dense_graph)
```
]
 
.right-panel-spatialmorphers2-user[
<PRE class="fansi fansi-output"><CODE># An sfnetwork with 1517 nodes and 1692 edges
#
# CRS: EPSG:4326 
#
# A directed simple graph with 1 component with spatially explicit edges
#
# Node Data: 1,517 x 1 (active)
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 geometry
 &lt;POINT [°]&gt;
1 (7.533722 51.95556)
2 (7.533461 51.95576)
3 (7.53236 51.95377)
4 (7.53209 51.95328)
5 (7.532442 51.95422)
6 (7.532823 51.95239)
# ... with 1,511 more rows
#
# Edge Data: 1,692 x 6
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 7.522594 ymin: 51.94151 xmax: 7.546705 ymax:
# 51.9612
 from to name type .tidygraph_edge~ geometry
 &lt;int&gt; &lt;int&gt; &lt;fct&gt; &lt;fct&gt; &lt;int&gt; &lt;LINESTRING [°]&gt;
1 1 2 Havix~ resid~ 1 (7.533722 51.95556, 7.53~
2 3 4 Piene~ secon~ 2 (7.53236 51.95377, 7.532~
3 5 3 Piene~ secon~ 2 (7.532442 51.95422, 7.53~
# ... with 1,689 more rows
</CODE></PRE>
]

---
class: split-40
count: false

```r
net %>% 
 convert(to_spatial_dense_graph) %>% 
* plot()
```
]
 
.right-panel-spatialmorphers2-user[
<img src="slides_files/figure-html/spatialmorphers2_user_2_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
*net_proj %>%
* activate("edges") %>%
* mutate(weight = edge_length()) %>%
* convert(to_spatial_shortest_paths, p3, p4)
```
]
 
.right-panel-spatialmorphers3-user[

```
# An sfnetwork with 45 nodes and 44 edges
#
# CRS: EPSG:3035 
#
# An unrooted tree with spatially explicit edges
#
# Edge Data: 44 x 7 (active)
# Geometry type: LINESTRING
# Dimension: XY
# Bounding box: xmin: 4150901 ymin: 3206483 xmax: 4152196 ymax:
# 3208337
 from to name type geometry weight
 <int> <int> <fct> <fct> <LINESTRING [m]> [m]
1 1 2 Pien~ seco~ (4151398 3207777, 415139~ 107.714~
2 3 4 Bred~ resi~ (4151845 3208048, 415192~ 100.608~
3 4 8 Bred~ resi~ (4152058 3208166, 415205~ 143.449~
4 9 10 <NA> cycl~ (4151674 3207951, 415167~ 59.453~
5 9 11 <NA> cycl~ (4151674 3207951, 415170~ 86.038~
6 13 14 Lise~ resi~ (4151129 3206483, 415124~ 112.524~
# ... with 38 more rows, and 1 more variable:
# .tidygraph_edge_index <int>
#
# Node Data: 45 x 2
# Geometry type: POINT
# Dimension: XY
# Bounding box: xmin: 4150901 ymin: 3206483 xmax: 4152196 ymax:
# 3208337
 geometry .tidygraph_node_index
 <POINT [m]> <int>
1 (4151398 3207777) 3
2 (4151370 3207673) 4
3 (4151845 3208048) 28
# ... with 42 more rows
```
]

---
class: split-40
count: false

```r
net_proj %>%  
  activate("edges") %>%  
  mutate(weight = edge_length()) %>%  
  convert(to_spatial_shortest_paths, p3, p4) ->  
* net_sp

*par(mar = c(1,1,1,1), bg = NA)
*plot(net_proj)
```
]
 
.right-panel-spatialmorphers3-user[
<img src="slides_files/figure-html/spatialmorphers3_user_2_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
net_proj %>%  
  activate("edges") %>%  
  mutate(weight = edge_length()) %>%  
  convert(to_spatial_shortest_paths, p3, p4) ->  
  net_sp

par(mar = c(1,1,1,1), bg = NA) 
plot(net_proj) 
*plot(net_sp,
* col = "Orange", lwd = 1.5, cex = 1.5,
* add = T)
```
]
 
.right-panel-spatialmorphers3-user[
<img src="slides_files/figure-html/spatialmorphers3_user_3_output-1.png" width="432" />
]

---

```r
*net_proj %>%
* morph(to_spatial_shortest_paths, p1, c(p3, p4))
```
]
 
.right-panel-spatialmorphers4-user[

```
# A tbl_graph temporarily morphed to a .f representation
# 
# Original graph is an undirected multigraph with 14 components
# consisting of 701 nodes and 851 edges
```
]

---
class: split-40
count: false

```r
net_proj %>%  
  morph(to_spatial_shortest_paths, p1, c(p3, p4)) %>%  
* activate("edges") %>%
* sapply(function(x) sum(st_length(x)))
```
]
 
.right-panel-spatialmorphers4-user[

```
[1] 1239.095 2651.979
```
]

---

```r
*ps1 = c(st_geometry(p1), st_sfc(p3))
*ps2 = c(st_geometry(p2), st_sfc(p4))
```
]
 
.right-panel-distancematrix-user[

]

---
class: split-40
count: false

```r
ps1 = c(st_geometry(p1), st_sfc(p3))  
ps2 = c(st_geometry(p2), st_sfc(p4))

*st_network_distance(net_proj, ps1, ps2)
```
]
 
.right-panel-distancematrix-user[

```
     [,1] [,2]
[1,]   16   28
[2,]   23   35
```
]

---
## Want more details?

Check the package vignettes:

- [Tidy Geospatial Networks in R](https://luukvdmeer.github.io/sfnetworks/articles/intro.html)

- [Extending the vocabulary](https://luukvdmeer.github.io/sfnetworks/articles/extensions.html)
---
background-image: url(https://pbs.twimg.com/media/EZvQNJlWoAE4D_c?format=jpg&name=4096x4096)

# Applications

Many potential applications including:

Transport planning/modelling (Robin), Road safety (Andrea), River network analysis (e.g. Water quality research, Flooding research, Aquatic ecology) Electricity network analysis, Social network analysis, Many More!

---

### How to cope with reduced public transport capacity post COVID-19?

Plans from Leeds City Council responding to national [guidance](https://www.gov.uk/government/publications/reallocating-road-space-in-response-to-covid-19-statutory-guidance-for-local-authorities) and [funding](https://www.gov.uk/government/news/2-billion-package-to-create-new-era-for-cycling-and-walking) for 'pop-up' cycleways:

![](https://pbs.twimg.com/media/EZ_-A0dXgAAlBzt?format=png&name=900x900)

---

## Spatial networks can support planning

### Transport networks are spatial networks

Spatial network representation in R:

![](https://raw.githubusercontent.com/cyipt/popupCycleways/master/figures/mapout.png)

---

background-image: url("https://raw.githubusercontent.com/cyipt/popupCycleways/master/figures/facet-output.png")
background-position: right
background-size: contain

### sf + igraph used in project to support investment in 'pop-ups' nationwide

Every city in England

Results in Leeds:

```
             name
1 Headingley Lane
2            A660
3  Woodhouse Lane
4             A65
5  Kirkstall Road
```

Cycleway planned for

Kirkstall Road (source: [BBC](https://www.bbc.co.uk/news/uk-england-leeds-52577554))

Free and open results:

[www.cyipt.bike/rapid](https://www.cyipt.bike/rapid)

Source code: [github.com/cyipt/popupCycleways](https://github.com/cyipt/popupCycleways)

---

## Ingredients to prioritise transport infrastructure

- Current travel patterns
- Road attribute data 
- Potential usage levels
- Impact on overall network

**Spatial networks can help, especially with the last two**

Source: https://www.pct.bike/

]

![](figs/pct.png)

]

---

## Why not use `stplanr::SpatialLinesNetwork`?

- You can:

```r
r_stplanr = stplanr::SpatialLinesNetwork(roxel)
```

```r
l = stplanr::route_local(r_stplanr, c(7.53, 51.95), c(7.54, 51.95))
plot(r_stplanr@sl$geometry)
plot(l$geometry, col = "red", lwd = 5, add = TRUE)
```

<img src="slides_files/figure-html/unnamed-chunk-8-1.png" width="432" />
]

Issues with stplanr implementation:

- `igraph` and `sf` parts are separate
- Hard to access the nodes
- Subsetting requires rebuilding the graph
- Cannot modify CRS 
- It's difficult to perform spatial or regular joins

```r
class(r_stplanr)
```

```
[1] "sfNetwork"
attr(,"package")
[1] "stplanr"
```

```r
slotNames(r_stplanr) # s4 class system
```

```
[1] "sl"          "g"           "nb"          "weightfield"
```

]

---

##  Another example: formatting `sfnetwork` data for a road safety model

We can use `sf`, `tidygraph`, `igraph` and several other packages for:

- summarizing the spatial dimension of the road network using several metrics (`tidygraph`);

- extracting the adjacency matrix of the edges, which is typically one of the ingredients in a spatial model (`igraph`);

- subsetting the road network considering only a particular geographical area (`sf`).

Let's see an example but first we need to create the `sfnetwork` version of `roxel` data:

```r
net <- as_sfnetwork(roxel, directed = FALSE)
```

---
class: center, middle

### Summarizing Graph Characteristics

---

```r
# Estimate nodes centrality degree
*net <- net
```
]
 
.right-panel-nodes_degree-user[

]

---
class: split-40
count: false

```r
# Estimate nodes centrality degree
net <- net %>% 
* activate("nodes")
```
]
 
.right-panel-nodes_degree-user[

]

---
class: split-40
count: false

```r
# Estimate nodes centrality degree
net <- net %>% 
 activate("nodes") %>% 
* mutate(degree = centrality_degree())
```
]
 
.right-panel-nodes_degree-user[

]

---
class: split-40
count: false

```r
# Estimate nodes centrality degree
net <- net %>% 
 activate("nodes") %>% 
 mutate(degree = centrality_degree())

# Plot with tmap. We need to escape the nodes and
# edges dimension and rebuild the sf structure.
*tm_shape(net %>% activate("edges") %>% st_as_sf()) +
* tm_lines()
```
]
 
.right-panel-nodes_degree-user[
<img src="slides_files/figure-html/nodes_degree_user_4_output-1.png" width="432" />
]

---
class: split-40
count: false

```r
# Estimate nodes centrality degree
net <- net %>% 
 activate("nodes") %>% 
 mutate(degree = centrality_degree())

# Plot with tmap. We need to escape the nodes and
# edges dimension and rebuild the sf structure.
tm_shape(net %>% activate("edges") %>% st_as_sf()) + 
 tm_lines() + 
*tm_shape(net %>% activate("nodes") %>% st_as_sf()) +
* tm_dots(
* size = 0.2,
* col = "degree",
* as.count = TRUE,
* palette = "viridis",
* title = ""
* ) +
*tm_layout(legend.text.size = 2)
```
]
 
.right-panel-nodes_degree-user[
<img src="slides_files/figure-html/nodes_degree_user_5_output-1.png" width="432" />
]

---

```r
# Estimate edges centrality betweenness
*net <- net %>%
* activate("edges")
```
]
 
.right-panel-edge_betweenness-user[

]

---
class: split-40
count: false

```r
# Estimate edges centrality betweenness
net <- net %>% 
 activate("edges") %>% 
* mutate(
* lengths = st_length(.),
* betweenness = centrality_edge_betweenness(lengths)
* )
```
]
 
.right-panel-edge_betweenness-user[

]

---
class: split-40
count: false

```r
# Estimate edges centrality betweenness
net <- net %>% 
 activate("edges") %>% 
 mutate( 
 lengths = st_length(.), 
 betweenness = centrality_edge_betweenness(lengths) 
 )

# Plot with tmap. We need to escape the nodes and
# edges dimension and rebuild the sf structure.
*tm_shape(net %>% activate("edges") %>% st_as_sf()) +
* tm_lines(
* n = 8,
* scale = 2,
* col = "betweenness",
* palette = "viridis",
* legend.col.show = FALSE
* ) +
*tm_shape(net %>% activate("nodes") %>% st_as_sf()) +
* tm_dots(size = 0.1)
```
]
 
.right-panel-edge_betweenness-user[
<img src="slides_files/figure-html/edge_betweenness_user_3_output-1.png" width="432" />
]

---

### Extracting the Adjacency Matrix of the edges

We can use `igraph` and `make_line_graph`! The `line graph` of a network object is obtained by inverting the nodes with the edges. We can apply this powerful idea to extract and visualize the `first-order binary adjacency matrix of the edges`. This type of matrix is used by a CAR or BYM spatial model.

---

class: split-50
count: false
 
###First order binary adjacency
.left-panel-line_graph-user[

```r
# Create the line graph
*net_line_graph <- make_line_graph(net)
```
]
 
.right-panel-line_graph-user[

]

---
class: split-50
count: false
 
###First order binary adjacency
.left-panel-line_graph-user[

```r
# Create the line graph
net_line_graph <- make_line_graph(net)

# Extract the first order binary adjacency matrix
*first_order_binary_adjacency <- net_line_graph %>%
* as_adjacency_matrix()
```
]
 
.right-panel-line_graph-user[

]

---
class: split-50
count: false
 
###First order binary adjacency
.left-panel-line_graph-user[

```r
# Create the line graph
net_line_graph <- make_line_graph(net)

# Extract the first order binary adjacency matrix
first_order_binary_adjacency <- net_line_graph %>% 
 as_adjacency_matrix()

*image(
* first_order_binary_adjacency[sample(1:851), ],
* xlab = list(label = "Columns", cex = 3),
* ylab = list(label = "Row", cex = 3),
* scales = list(cex = 3),
* lwd = 3,
* sub = ""
* )
```
]
 
.right-panel-line_graph-user[
<img src="slides_files/figure-html/line_graph_user_3_output-1.png" width="432" />
]

---

class: split-50
count: false
 
###Second order binary adjacency
.left-panel-line_graph2-user[

```r
# Similar idea for second order binary adjacency
*ego2_net <- ego(net_line_graph, 2)
```
]
 
.right-panel-line_graph2-user[

]

---
class: split-50
count: false
 
###Second order binary adjacency
.left-panel-line_graph2-user[

```r
# Similar idea for second order binary adjacency
ego2_net <- ego(net_line_graph, 2)

*second_order_binary_adjacency <- ego2_net %>%
* graph_from_adj_list() %>%
* as_adjacency_matrix()
```
]
 
.right-panel-line_graph2-user[

]

---
class: split-50
count: false
 
###Second order binary adjacency
.left-panel-line_graph2-user[

```r
# Similar idea for second order binary adjacency
ego2_net <- ego(net_line_graph, 2)

second_order_binary_adjacency <- ego2_net %>% 
 graph_from_adj_list() %>% 
 as_adjacency_matrix()

*image(
* second_order_binary_adjacency[sample(1:851), ],
* xlab = list(label = "Columns", cex = 3),
* ylab = list(label = "Row", cex = 3),
* scales = list(cex = 3),
* lwd = 3,
* sub = ""
* )
```
]
 
.right-panel-line_graph2-user[
<img src="slides_files/figure-html/line_graph2_user_3_output-1.png" width="432" />
]

---
class: center, middle

### Thank you for you interest!

#### Questions?

.note[Slides powered by [xaringan](https://github.com/yihui/xaringan), [xaringanthemer](https://github.com/gadenbuie/xaringanthemer) and [flipbookr](https://github.com/EvaMaeRey/flipbookr)]

---
background-image: url(https://github.com/sfnetworks/sfnetworks-webinar/raw/master/figs/e-Rum.png)

## See you tomorrow at e-Rum2020!