Maps & Nets

.title[
# Maps & Nets
]
.author[
### Visualizing and communicating data with R <a href='https://therbootcamp.github.io'> The R Bootcamp @ Unibas </a> <a href='https://therbootcamp.github.io/DataViz_Unibas2023Nov/'> </a>  <a href='https://therbootcamp.github.io'> </a>  <a href='therbootcamp@gmail.com'> </a>  <a href='https://www.linkedin.com/company/27090302'> </a>
]
.date[
### November 2023
]

---

<div class="my-footer">
 
 
 <img src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/by-sa.png" height=14 style="vertical-align: middle"/>
 
 <a href="https://therbootcamp.github.io/">
 
 
 https://therbootcamp.github.io
 
 
 </a>
 <a href="https://therbootcamp.github.io/">
 
 The R Bootcamp | November 2023
 
 </a>
 
 </div>

---

# Maps with <mono>ggplot</mono>/<mono>sf</mono>

<ul>
 <li class="m1">Maps require geometric shapes stored in <high>shapefiles</high>.</li> 
 <li class="m2">The <high>simple features</high> (<mono>sf</mono>) framework makes processing and visualizing maps with the <mono>tidyverse<mono> easy.</li>
</ul>

]

]

---

# Shapefiles

<ul>
 <li class="m1">Geospatial vector data format for <high>geographic information system</high> (GIS) software.</li> 
 <li class="m2">Necessary files:</li>
 <ul class="level">
 <li><mono>.shp</mono> | actual shapefile</li>
 <li><mono>.shx</mono> | shape index format</li>
 <li><mono>.dbf</mono> | attribute format</li>
 </ul> 
 <li class="m3">Optional files:</li>
 <ul class="level">
 <li><mono>.prj</mono> | projection description</li>
 <li><mono>.cpg</mono> | code page specification</li>
 </ul>
</ul>

]

---

# <mono>sf</mono>

```r
read_sf('1_Data/quarters') 
```

```
Simple feature collection with 21 features and 5 fields
Geometry type: POLYGON
Dimension: XY
Bounding box: xmin: 2609000 ymin: 1263000 xmax: 2619000 ymax: 1272000
Projected CRS: CH1903+ / LV95
# A tibble: 21 × 6
 OBJID OBJECTID TXT ZTXT TYPE geometry
 <chr> <dbl> <chr> <chr> <chr> <POLYGON [m]>
 1 17136 1 7 07 Bruderholz ((2612556 1264548, 2612561 1264514, 2612565 12644…
 2 17139 2 6 06 Gundeldingen ((2610887 1266551, 2610896 1266546, 2610918 12665…
 3 17142 3 5 05 St. Alban ((2612942 1267023, 2613000 1267019, 2613027 12670…
 4 17145 4 4 04 Breite ((2613684 1266891, 2613686 1266889, 2613689 12668…
 5 17148 5 8 08 Bachletten ((2610561 1266791, 2610571 1266781, 2610595 12667…
 6 17151 6 2 02 Vorstädte ((2610928 1268323, 2610934 1268293, 2610947 12682…
 7 17154 7 1 01 Altstadt Grossbasel ((2611366 1267578, 2611375 1267568, 2611376 12675…
 8 17157 8 3 03 Am Ring ((2610705 1267923, 2610680 1267844, 2610647 12677…
 9 17160 9 9 09 Gotthelf ((2609245 1266802, 2609236 1266801, 2609366 12671…
10 17163 10 10 10 Iselin ((2610234 1267656, 2610284 1267389, 2610234 12674…
# ℹ 11 more rows
```

---

# <mono>geom_sf</mono>

<ul>
 <li class="m1">Since <mono>read_sf</mono> creates a <mono>tibble</tibble>, it can be plugged <high>straight into <mono>ggplot</mono></high>.</li> 
 <li class="m2">The dedicated geom <mono>geom_sf</mono> plots the geometric polygons.</li>
</ul>

```r
# read shapefiles
quarters_map <- read_sf('1_Data/quarters')

# plot quarters
quarters_map %>% 
  ggplot() +
  geom_sf()
```

]

]

---

# <mono>geom_sf</mono>

```r
# read shapefiles
quarters_map <- read_sf('1_Data/quarters')

# plot quarters
quarters_map %>% 
  ggplot() +
  geom_sf() + 
  
  # remove background
  theme_void()
```

]

]

---

# Wrangling

<ul>
 <li class="m1">Since <mono>read_sf</mono> creates a <mono>tibble</tibble>, one can easily <high>join additional data</high>.</li>
</ul>

```r
# join basel tax data
quarters_map <- quarters_map %>% 
 left_join(filter(basel, year == 2017), 
 by = c("TYPE" = "quarter"))

# plot quarters
quarters_map %>% 
  ggplot() +
  geom_sf() + 
  theme_void()
```

]

]

---

# Add color

<ul>
 <li class="m1">Colors are set using <mono>aes()</mono> just like in a regular <mono>ggplot</mono>.</li>
</ul>

```r
# join basel tax data
quarters_map <- quarters_map %>% 
 left_join(filter(basel, year == 2017), 
 by = c("TYPE" = "quarter"))

# plot quarters
quarters_map %>% 
  ggplot() +
  
  # fill color by income
  geom_sf(aes(fill = income_mean)) + 
  theme_void()
```

]

]

---

# Styling

<ul>
 <li class="m1">Styling can be adjusted just like in any other <mono>ggplot<mono>.</li>
</ul>

```r
# join basel tax data
quarters_map <- quarters_map %>% 
 left_join(filter(basel, year == 2017), 
 by = c("TYPE" = "quarter"))

# plot quarters
quarters_map %>% 
  ggplot() +
  
  # add white outlines
  geom_sf(aes(fill = income_mean),
              col = "white") + 
  theme_void()
```

]

]

---

# Styling

<ul>
 <li class="m1">Styling can be adjusted just like in any other <mono>ggplot<mono>.</li>
</ul>

```r
# join basel tax data
quarters_map <- quarters_map %>% 
 left_join(filter(basel, year == 2017), 
 by = c("TYPE" = "quarter"))

# plot quarters
quarters_map %>% 
  ggplot() +
  geom_sf(aes(fill = income_mean),
              col = "white") + 
  theme_void() + 
  
  # change legend title
  scale_fill_continuous(name = 'Income') 
```

]

]

---

# Styling

<ul>
 <li class="m1">Styling can be adjusted just like in any other <mono>ggplot<mono>.</li>
</ul>

```r
# join basel tax data
quarters_map <- quarters_map %>% 
 left_join(filter(basel, year == 2017), 
 by = c("TYPE" = "quarter"))

# plot quarters
quarters_map %>% 
  ggplot() +
  geom_sf(aes(fill = income_mean),
              col = "white") + 
  theme_void() + 
  scale_fill_continuous(name = 'Income') +
  
  # add annotion
  labs(title = "Inequality in Basel",
     subtitle = "Average income in Basel...",
     caption = "ource: Open Data Basel...")
```

]

]

---

# <mono>geom_label()</mono>

<ul>
 <li class="m1">Annotation with labels or images requires specification of a location such as the <high>polygon's centroid</high>.</li>
</ul>

```r
# start with joint object
quarters_map %>% 
  
  # determine centroids
  mutate(ctr = st_centroid(geometry),
    lon = sapply(ctr, function(x) x[1]),
    lat = sapply(ctr, function(x) x[2])) %>%

# plot quarters
  ggplot(...) +
    
  # add labels
  geom_label_repel(aes(x = lon, 
                       y = lat, 
                       label = TYPE), 
                   size = 2.5)
```

]

]

---

# <mono>geom_image()</mono>

<ul>
 <li class="m1">Styling can be adjusted just like in any other <mono>ggplot<mono>.</li>
</ul>

```r
# join basel tax data
quarters_map %>% 
  
  # determine centroids
  mutate(ctr = st_centroid(geometry),
    lon = sapply(ctr, function(x) x[1]),
    lat = sapply(ctr, function(x) x[2])) %>%

# plot quarters
  ggplot(...) +
    
  # add images
  geom_image(aes(x = lon, y = lat, 
                 image = image), 
             size=.05)
```

]

]

---

# Networks with <mono>tidygraph</mono> & <mono>ggraph</mono>

<ul>
 <li class="m1">Networks (graphs) show relationships (edges) between units (nodes).</li> 
 <li class="m2"><highm>tidygraph</highm> helps with creating tidy network objects, <highm>ggraph</highm> helps with visualizing networks within the <mono>ggplot</mono> framework.</li>
</ul>

]

]

---

# Edge list

<ul>
 <li class="m1">A network is typically created from either an <high>edge list</high> or an adjacency matrix.</li> 
</ul>

```r
# determine correlations
basel %>% 
 select(year, quarter, income_median) %>% 
 pivot_wider(names_from = quarter, 
 values_from = income_median) %>% 
 select(-1) %>% 
 cor() %>% 
 as.table() %>% 
 as_tibble(.name_repair = "unique") %>% 
 rename(from = `...1`,
 to = `...2`,
 weight = n) %>% 
 
 # limit to strong edges
 filter(weight < 1, weight > .5)
```

]

```
# A tibble: 184 × 3
 from to weight
 <chr> <chr> <dbl>
 1 Am Ring Altstadt Grossbasel 0.523
 2 Bachletten Altstadt Grossbasel 0.558
 3 Am Ring Vorstädte 0.881
 4 St. Alban Vorstädte 0.806
 5 Gundeldingen Vorstädte 0.721
 6 Bruderholz Vorstädte 0.605
 7 Bachletten Vorstädte 0.770
 8 Gotthelf Vorstädte 0.872
 9 Altstadt Kleinbasel Vorstädte 0.560
10 Wettstein Vorstädte 0.917
# ℹ 174 more rows
```

]

---

# <mono>tidygraph</mono>

<ul>
 <li class="m1"><mono>tidygraph</mono> generates a <high>network object</high> the edge list.</li> 
</ul>

```r
# determine correlations
network <- basel %>% 
 select(year, quarter, income_median) %>% 
 pivot_wider(names_from = quarter, 
 values_from = income_median) %>% 
 select(-1) %>% 
 cor() %>% 
 as.table() %>% 
 as_tibble(.name_repair = "unique") %>% 
 rename(from = `...1`,
 to = `...2`,
 weight = n) %>% 
 filter(weight < 1, weight > .5) %>% 
 
 # format to graph
 as_tbl_graph(edges, directed = FALSE)
```

]

```
# A tbl_graph: 21 nodes and 184 edges
#
# An undirected multigraph with 1 component
#
# A tibble: 21 × 1
 name 
 <chr> 
1 Am Ring 
2 Bachletten 
3 St. Alban 
4 Gundeldingen
5 Bruderholz 
6 Gotthelf 
# ℹ 15 more rows
#
# A tibble: 184 × 3
 from to weight
 <int> <int> <dbl>
1 1 13 0.523
2 2 13 0.558
3 1 14 0.881
# ℹ 181 more rows
```

]

---

# <mono>ggraph</mono>

<ul>
 <li class="m1"><mono>ggraph</mono> provides syntax to plot networks within the <highm>ggplot framwork</high>.</li> 
</ul>

```r
# start network plot
network %>% 
  ggraph() 
```

]

]

---

# Adding nodes & edges

<ul>
 <li class="m1"><mono>geom_node_point()</mono> adds points to represent the nodes of the network.</li> 
 <li class="m2"><mono>geom_edge_link()</mono> adds (straight) lines to represent the edges of the network.</li> 
</ul>

```r
# start network plot
network %>% 
  ggraph() + 
  
  # add nodes and edges
  geom_edge_link() +
  geom_node_point()
```

]

]

---

# Adding labels

<ul>
 <li class="m1"><mono>geom_node_label()</mono> adds node labels.</li> 
</ul>

```r
# start network plot
network %>% 
  ggraph() + 
  geom_edge_link() +
  geom_node_point() +

# add labels
  geom_node_label(aes(label = name), 
                  show.legend = FALSE) 
```

]

]

---

# Styling networks

<ul>
 <li class="m1"><mono>theme_graph()</mono> removes unnecessary background details.</li> 
 <li class="m2"><mono>scale_edge_width()</mono> and similar functions can be used to adapt edges or nodes.</li>
</ul>

```r
# start network plot
network %>% 
  ggraph() + 
  geom_edge_hive() + # style change
  geom_node_point() +
  geom_node_label(aes(label = name), 
                  show.legend = FALSE) +
  
  # add styling
  theme_graph() + 
  scale_edge_width(range=c(.2, 1)) +
  coord_cartesian(clip = "off")
```

]

]

---

# Adding clusters

<ul>
 <li class="m1"><mono>group_louvain()</mono> and similar functions can be used to determine node groups (clusters).</li>
</ul>

```r
# add clusters
network %>% 
  mutate(community = 
           group_louvain(weight) %>% 
           as.factor()) %>%
  
  # network plot
  ggraph() + 
  geom_edge_hive() +
  geom_node_point() +
  geom_node_label(aes(label = name,
                      col = community), 
                  show.legend = FALSE) +
  theme_graph() + 
  scale_edge_width(range=c(.2, 1)) +
  coord_cartesian(clip = "off")
  scale_color_manual(values=viridis(3)[1:2])
```

]

]

---

# Adding ellipses

<ul>
 <li class="m1"><mono>stat_ellipse()</mono> can be used to add ellipses covering points in the data (similar <mono>stat_*</mono> functions exist for other visual elements.</li>
</ul>

```r
# add clusters
network %>% 
  ...
  
  # network plot
  ggraph() + 
  
  # ellipses
  stat_ellipse(aes(x = x , y = y, 
                   group = community, 
                   fill = community), 
               level = .8, geom = "polygon", 
               alpha = .1, show.legend = FALSE) + 
  
  # rest of plot
  scale_fill_manual(values=viridis(3)[1:2]) +
  ...
```

]

]
---

# Adding annotation

<ul>
 <li class="m1"><mono>annotate()</mono> can be used to <high>add any type of plot elements</high> (layers) using vector specification.</li> 
</ul>

```r
# add clusters
network %>% 
  ...
  
  # network plot
  ggraph() + 
  ...
  
  # annotation
  annotate("text", 
           x = c(-2.5, 2), y = c(.8,1), 
           label = c("Group 1","Group 2"), 
           size=7)
```

]

]

---

# Adding labels

<ul>
 <li class="m1">Labels can be added as usual using <mono>labs()</mono>.</li> 
</ul>

```r
# add clusters
network %>% 
  ... 
  
  # network plot
  ggraph() + 
  ...
  
  # labels
  labs(title = "Network...",
       subtitle = "reveals...")
```

]

]

---

<h1>click to download <a href="https://www.dropbox.com/s/mx2x5bak5gie7sb/Maps%26Nets.Rmd?dl=1">Practical</a></h1>