07.Rmd

# Network Visualization and Aesthetics

**Learning objectives:**

-   How to change the aesthetics of network visualizations.

## The Basics

```{r}
suppressPackageStartupMessages(library(igraph))
suppressPackageStartupMessages(library(tidygraph))
suppressPackageStartupMessages(library(ggraph))
suppressPackageStartupMessages(library(tidyverse))
```

Visualization and statistics are the primary tools at our disposable for conveying convincing stories about the structure and dynamics of the networks we study.

Load data and create an igraph object from edgelist

```{r}
#read data
data_url <- "https://raw.githubusercontent.com/mahoffman/social_network_analysis/master/Data/money_edgelist.csv"
money_edgelist <- read_csv(data_url)
#igraph
moneyNetwork_ig <- money_edgelist %>% 
  as.matrix() %>% 
  graph_from_edgelist(directed=TRUE)
#tidygraph
moneyNetwork_tg <- as_tbl_graph(moneyNetwork_ig)
```

```{r}
money_edgelist
moneyNetwork_ig
moneyNetwork_tg
```

Basic igraph plot

```{r}
set.seed(1992)
plot(moneyNetwork_ig)
```

```{r}
set.seed(1992)
plot(moneyNetwork_ig)
```

Save "nice" layout

```{r}
nt_layout <- layout_nicely(moneyNetwork_ig)
```

Set the layout

```{r}
plot(moneyNetwork_ig, layout = nt_layout)
```

Basic ggraph plot

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link() + 
  geom_node_point()
```

### Nodes

We can control the size of the nodes

```{r}
plot(moneyNetwork_ig, layout = nt_layout, vertex.size = 50)
plot(moneyNetwork_ig, layout = nt_layout, vertex.size = 10)
```

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link() + 
  geom_node_point(size = 5)
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link() + 
  geom_node_point(size = 10)
```

We can also control the color of the nodes and their frames

```{r}
plot(moneyNetwork_ig,
     layout = nt_layout,
     vertex.size = 10,
     vertex.color = "tomato",
     vertex.frame.color = NA #remove the frames of the nodes
     )
```

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link() + 
  geom_node_point(color = "tomato", size = 5)
```

Adjust label size with vertex.label.cex. We can adjust the color with vertex.label.color

```{r}
plot(moneyNetwork_ig,
     layout = nt_layout,
     vertex.size = 10,
     vertex.color = "tomato",
     vertex.frame.color = NA,
     vertex.label.cex = .7, #adjust label size
     vertex.label.color = "black" #adjust label color
     )
```

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link() + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

Alternatively, if we want to get rid of the labels, we can just set vertex.label to NA.

```{r}
plot(moneyNetwork_ig,
     layout = nt_layout,
     vertex.size = 10,
     vertex.color = "tomato",
     vertex.frame.color = NA,
     vertex.label = NA#remove label
     )
```

## Edges

Adjust edges size and curvature to give them a nicer aesthetic

```{r}
plot(moneyNetwork_ig,
     layout = nt_layout,
     vertex.size = 10,
     vertex.color = "tomato",
     vertex.frame.color = NA,
     vertex.label.cex = .7,
     vertex.label = NA,
     edge.curved = 0.1, #curvature of the arrow(higher value = more curvature; 0=no curvature)
     edge.arrow.size = 0.3, #size of arrow head
     edge.width = .7 #width of the edge
     )
```

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_arc(strength = 0.1) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))

ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_bend(strength = 0.2) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))

ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_diagonal(strength = 0.7) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

```{r}
ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_link(aes(start_cap = label_rect(node1.name),
                     end_cap = label_rect(node2.name)),
                 color = "grey60",
                 strength = 0.1,
                 width = 1,
                 arrow = arrow(angle = 15,
                               length=unit(0.5,"cm"),
                               ends="first",
                               type = "closed")) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

```{r}

ggraph(moneyNetwork_tg, layout = nt_layout) + 
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

for more tricks check this notebook <https://rpubs.com/jmmcclu3/865821>

## Layouts

![source: https://www.data-imaginist.com/2017/ggraph-introduction-layouts/](https://d33wubrfki0l68.cloudfront.net/756daa22f3690a4073ce41607ef98bf5536035a4/1f06e/post/2017-02-06-ggraph-introduction-layouts_files/figure-html/unnamed-chunk-12.gif)

-- A network's layout determines the nodes' positions in the plot. -- All layouts try to minimize the number of edges that cross.

The author states that he usually uses either the Kamada Kawai (`igraph::layout.kamada.kawai()`; `ggraph:::layout_tbl_graph_igraph()`) algorithm or the Fruchterman Reingold algorithm.

**Kamada Kawai**

```{r}
# first we run the layout function on our graph
kamadaLayout <- layout.kamada.kawai(moneyNetwork_ig)

# and then we change the default layout setting to equal the layout we generated above
plot(moneyNetwork_ig,
     layout = kamadaLayout,
     vertex.size = 10,
     vertex.color = "tomato",
     vertex.frame.color = NA,
     vertex.label.cex = .7,
     vertex.label = NA,
     edge.curved = .1,
     edge.arrow.size = .3,
     edge.width = .7)
```

```{r}
ggraph(moneyNetwork_tg, layout = kamadaLayout) + 
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

The layout can be directly stated in ggraph call

```{r}
ggraph(moneyNetwork_tg, layout = "kk") + #kamada.kawai
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(color = "tomato", size = 5)+
  geom_node_text(aes(label = name))
```

Check for layouts available ggraph from igraph ?ggraph:::layout_tbl_graph_igraph()

## Adding attributes to a network object

Load node attributes (metadata)

```{r}
data_url2 <- "https://raw.githubusercontent.com/mahoffman/social_network_analysis/master/Data/attribute_df.csv"
attributes <- read_csv(data_url2)
head(attributes)
```

add attributes to the igraph object

```{r}
moneyNetwork_ig <- graph_from_data_frame(money_edgelist,
                                      directed = T,
                                      vertices = attributes)
moneyNetwork_ig

V(moneyNetwork_ig)$Gender 
```

and the ggraph object

```{r}
moneyNetwork_tg <- tbl_graph(edges = money_edgelist,
                             directed = T,
                             nodes = attributes)
moneyNetwork_tg
```

Let's set a layout for the network

```{r}
kk_layout <- layout.kamada.kawai(moneyNetwork_ig)
```

## Plotting based on attributes

```{r}
V(moneyNetwork_ig)$color <- ifelse(V(moneyNetwork_ig)$Gender == "Male", "dodgerblue3","seagreen")
```

```{r}
plot(moneyNetwork_ig,
     layout = kk_layout,
     vertex.size = 10,
     vertex.frame.color = "black", 
     vertex.label.cex = .7, 
     vertex.label = NA, 
     edge.curved = .1, 
     edge.arrow.size = .3)
```

```{r}
ggraph(moneyNetwork_tg, layout = kk_layout) + #kamada.kawai
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(aes(color = Gender), size = 5)+
  scale_color_manual(values = c(Female = "seagreen", Male = "dodgerblue3"))
```

Now let's try role. First, I reset color. There are four roles (Father, Mother, Son, Daughter), so we need a few more ifelse statements to code for all of them.

```{r}
V(moneyNetwork_ig)$color <- NA
role_ig <- V(moneyNetwork_ig)$Role
V(moneyNetwork_ig)$color <- case_when(
 role_ig == "Father" ~ "burlywood1",
  role_ig == "Mother" ~ "seagreen",
  role_ig == "Son" ~ "grey70",
  TRUE ~ "tomato")

plot(moneyNetwork_ig,
     layout = kk_layout,
     vertex.size = 10,
     vertex.label.cex = .7,
     vertex.label = NA,
     edge.curved = .1, 
     vertex.frame.color = "black",
     edge.arrow.size = .3,
     edge.width = .7,
     edge.color = "grey30")
```

```{r}
ggraph(moneyNetwork_tg, layout = kk_layout) + #kamada.kawai
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(aes(color = Role), size = 5)+
  scale_color_manual(values = c(Father = "burlywood1",
                                Mother = "seagreen",
                                Son = "grey70",
                                Daughter = "tomato"))
```

Last but not least, let's adjust the sizes of the nodes so that they reflect differences in age. We can set node size to be 1/5th of the node's age with the code below. Simple, but effective. It looks like the oldest nodes give to the most people.

```{r}
V(moneyNetwork_ig)$size = V(moneyNetwork_ig)$Age/5

plot(moneyNetwork_ig, 
     layout = kk_layout,
     vertex.label.cex = .7,
     vertex.label = NA,
     edge.curved = .1,
     vertex.frame.color = "black",
     edge.arrow.size = .3, 
     edge.width = .7,
     edge.color = "grey30")
```
```{r}
ggraph(moneyNetwork_tg, layout = kk_layout) + #kamada.kawai
  geom_edge_fan(width = 1,
                aes(alpha = after_stat(index)),
                show.legend = FALSE) + 
  geom_node_point(aes(color = Role,
                      size = Age/5))+
  scale_color_manual(values = c(Father = "burlywood1",
                                Mother = "seagreen",
                                Son = "grey70",
                                Daughter = "tomato"))
```
## Meeting Videos

### Cohort 1

`r knitr::include_url("https://www.youtube.com/embed/H8ihh6AyZ4I")`

<details>

<summary>

Meeting chat log

</summary>
```
00:06:22	Abdalrhman Mostafa:	Hi Mohamed, ezayak!
00:09:24	mohamed.shoeb:	https://bookdown.org/markhoff/social_network_analysis/
```
</details>