Networks in Archaeology

class: center, middle, inverse, title-slide

.title[
# Networks in Archaeology
]
.author[
### 
]
.date[
### 2024/02/29
]

---

## Preamble

1. What is a 'network'

2. What are they useful for

3. How can they be applied in Archaeology
   
--

- Example
--

- Limits

--
 
.pull-left[

- https://cdal.arch.cam.ac.uk/A11-networks/
- https://cdal.arch.cam.ac.uk/A11-networks/presentation.html

]

.pull-right[
![](presentation_files/figure-html/unnamed-chunk-2-1.png)
]

---
class: inverse, center, middle

# What is a network?

---

### What is a network?

.pull-left[
> Networks are nothing more than a set of entities and the pairwise connections among them.

]

.pull-right[
<img src="images/archnetsci_bookcover.png" alt="CDAL" width="200px">

From Brughmans & Peeples 2023

]

.pull-left[
Ressources:

- https://archnetworks.net/
  ]

---

### What is a network?

.pull-left[

Sociology, since August Comte:

- Understanding human society through interactions between actors of the society

]

- From early sociologists at the end of XIXth early XXth following Comte: Durkheim, Lebon, Simmel, combined with biometrician to sociometrician (Moreno)

.pull-right[

<img src="images/linkBank.png" alt="CDAL" width="200px">
 
(From Hobson 1894, in Freeman 2003)

]

<img src="images/netfoot.png" alt="CDAL" width="150">
 
(From Moreno 1934, in Freeman 2003)

.pull-left[
- In the 90s, more data, more network

]

- A "science of Networks"

.pull-right[

<img src="images/networkAnIntro.webp" alt="CDAL" width="150px">
(Newman 2010)

]

---

### Formal Networks

Simple but very handy formalisation:

- Nodes (vertices/agents/actors/sites/...) : A, B, C

- Links (edges/connection): presence/absence (0-1, True, False) or weighted (distance,score,etc...).

- Edge list

.pull-left[

```r
A,C,1 # link between A & C
B,C,5 # strong link between B&C
```
]
.pull-right[

]

- Adjacency list

- Adjacency matrix....

.small-left[

```r
  A B C
A 0 0 1
B 0 0 5
C 0 0 0
```
]

---
class: inverse,

## Formal Network are easy to handle

## ...

## Even with R!

---

### Network with R

A few examples

(quick note)

**The igraph package:**

Throughout this session we will use the package igraph to do our analysis

```r
library(igraph)
```

There are plenty of other well known packages and program to do so such as:
--

- [python-igraph](https://python.igraph.org/): igraph in python (and other)
- [networkx](https://networkx.github.io/): another python package
- [Gephi](https://gephi.org/) a powerful and intuitive graphical interface

---

### Network with R

A few examples

.pull-left[

```r
A,C,1 # link between A & C
B,C,5 # strong link between B&C
```
Taking back the previous example

```r
library(igraph)

adjlist<-read.table(text="
 v1 v2 weight
 A C 1 # link between A & C
 B C 5 # strong link between B&C
",header=T,comment="#")

graph <- graph.data.frame(adjlist)
par(mar=c(0,0,0,0))
plot(graph)
mat=as_adjacency_matrix(graph,attr="weight")
```

]

```
## 3 x 3 sparse Matrix of class "dgCMatrix"
##   A B C
## A . . 1
## B . . 5
## C . . .
```

.pull-right[
![](presentation_files/figure-html/simptree-1.png)
]

---

### Network with R

A few examples

.pull-left[

```r
library(igraph)

adjlist<-read.table(text="
 v1 v2 weight
 A B 1 
 A C 2 
 C A 2 
 B C 8 
 D E 10 
 D A 1 
 D C 1 
",header=T,comment="#")

graph <- graph.data.frame(adjlist)
plot(graph)
mat=as_adjacency_matrix(graph,attr="weight")
```

]

```
## 5 x 5 sparse Matrix of class "dgCMatrix"
##   A C B D  E
## A . 2 1 .  .
## C 2 . . .  .
## B . 8 . .  .
## D 1 1 . . 10
## E . . . .  .
```

.pull-right[
![](presentation_files/figure-html/simptree2-1.png)
]

---

### Network with R

A few examples

.pull-left[

```r

adjlist<-read.table(text="
 ,DB,P,RM,G,CC 
BA,1,0,1,0,1
JL,1,0,0,0,1
MR,0,0,0,1,1
MSLL,1,0,0,0,1
NS,0,1,1,0,0
W,1,0,1,0,0
",header=T,sep=",")

aa=as.matrix(adjlist[,2:6])
rownames(aa)=adjlist[,1]

graph=graph_from_incidence_matrix(aa)
plot(graph,vertex.color=ifelse(V(graph)$type,1,2))
```

]

.pull-right[
![](presentation_files/figure-html/hobsonex-1.png)
]

---

### Network with R

```r
par(mfrow=c(1,2),mar=c(0,0,0,0))

plot(bipartite_projection(graph)$proj2,vertex.color=1)
plot(bipartite_projection(graph)$proj1,vertex.color=2)
```

![](presentation_files/figure-html/hobsonex2-1.png)

---

### Examples Everywhere

Where there are _things_ and _links_, there are networks.

.small-left[
Genetics

<a href="images/genetics.png">
<img src="images/genetics.png" alt="genet" width="300px">
</a>

(Atiia et al. 2020)
]

.small-left[
Biochemistry

<a href="images/biochemistry.png">
<img src="images/biochemistry.png" alt="biochem" width="300px">
</a>

(Ochieng et al. 2023)
]

.small-left[
Epidemiology

<a href="images/epidemio-net.png" >
<img src="images/epidemio-net.png" alt="epinet" width="300px">
</a>
(Silk et al. 2022)

]

.small-left[
Ecology 

<a href="images/eco-net.png">
<img src="images/eco-net.png" alt="ecology" width="70">
</a>

(Losapio et al. 2019)
]

.small-left[
Economy

<a href="images/politics.png" >
<img src="images/politics.png" alt="geopol" width="300px">
</a>

(Fagiolo & Luzzati 2023)
]

.small-left[
Sociology

<a href="images/socialnet.png" >
<img src="images/socialnet.png" alt="geopol" width="350px">
</a>

(Gaumont et al. 2018)
]

.small-left[
History

<video width="400" controls autoplay>
 <source src="images/1240064s2.mov" type="video/mp4">
</video>

(Schich et al. 2014)
]

---
class: inverse,center,middle

# Explore & Compare networks and their properties

---

### A whole family of network types & Metrics

.pull-left[

- Trees

]

- Random

- Small world

- Scale Free

- ...

.pull-right[

- In and out degree

]

- Shortest Paths

- 'Length'
{{content}}

- Clustering

- Centrality

- ...

---

### Full graph

A full graph is a graph where all nodes are connected

.pull-left[

```r
fg <-make_full_graph(50)
lfg=layout_nicely(fg)
plot(fg,vertex.size=3,vertex.label=NA,edge.width=.1,layout=lfg)
```

]

.pull-right[
![](presentation_files/figure-html/fullgraph-1.png)
]

---

### Random Graph

Lot of way to do a 'random' graph, here adding node and randomly adding a edge

.pull-left[

```r
rgg <- sample_growing(80, citation=FALSE,directed=FALSE)
rgg=simplify(delete.vertices(rgg,which(degree(rgg)==0)))
rggln=layout_nicely(rgg)

plot(rgg,vertex.size=3,vertex.label=NA,edge.width=1,layout=rggln)
```

]

.pull-right[
![](presentation_files/figure-html/randgraph-1.png)
]

---

### Tree

A tree is a directed graph with no loops. When no root is chose they are usually represented like this:

.pull-left[

```r
tr <-make_tree(50,children=3,mode="undirected")
plot(tr)
```

A graph made by 40 nodes, each node having maximum 3 children

]

.pull-right[
![](presentation_files/figure-html/tree_unrooted-1.png)
]

---

### Tree

A tree is a directed graph with no loops. But more often, a root is chosen, which will be use at the `top` of the `tree` (the `root` of  a rotated `tree`.

.pull-left[

```r
plot(tr,layout=layout_as_tree(tr,root=1))
```

]

.pull-right[
![](presentation_files/figure-html/tree_rooted-1.png)
]

---

### Small-world

Watts-Strogatz model creates a lattice (with `dim` dimensions and `size` nodes across dimension) and rewires edges randomly with probability `p`.

.pull-left[

```r
swA <- sample_smallworld(dim=2, size=7, nei=1, p=0.02)
plot(swA, vertex.size=degree(swA), vertex.label=NA, layout=layout_in_circle)
```

A 49 nodes graph connected as a "small world".

]

.pull-right[
![](presentation_files/figure-html/smallworldA-1.png)
]

---

### Small-world

.pull-left[

```r
sw <- sample_smallworld(dim=2, size=7, nei=1, p=0.1)
swlc=layout_in_circle(sw)
plot(sw, vertex.size=degree(sw), vertex.label=NA, layout=swlc)
```

A 49 nodes graph connected as a "small world".

]

.pull-right[
![](presentation_files/figure-html/smallworld-1.png)
]

---

### Scale-free graphs

Using Barabasi-Albert preferential attachment model
(`n` is number of nodes, `power` is the power of attachment (`1` is linear); `m` is the number of edges added on each time step)

.pull-left[

```r
ba <- sample_pa(n=50, power=1.01, m=1, directed=F)
baln=layout_nicely(ba)
plot(ba, vertex.size=degree(ba), vertex.label=NA,layout=baln)
```
]

.pull-right[
![](presentation_files/figure-html/scalefree-1.png) 
]

---

### Multiple metrics to describe them

- In and out degree
- Shortest Paths
- 'Length'
---

### In and Out Degree

.pull-left[

```r
node
```

```
## [1] 6
```

```r
degree(sw)[node]
```

```
## [1] 5
```

```r
plot(sw, layout=layout_in_circle)
```

]

.pull-right[

![](presentation_files/figure-html/degreeilu-1.png)

]

---

### In and Out Degree

.pull-left[

```r
dgr=degree(sw)
dgr=rbind("id"=seq_along(dgr),"degree"=dgr)
print(dgr)
```

```
##        [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9] [,10] [,11] [,12] [,13] [,14] [,15] [,16] [,17] [,18] [,19]
## id        1    2    3    4    5    6    7    8    9    10    11    12    13    14    15    16    17    18    19
## degree    3    3    4    5    5    5    3    2    3     3     3     6     4     4     4     4     4     4     3
##        [,20] [,21] [,22] [,23] [,24] [,25] [,26] [,27] [,28] [,29] [,30] [,31] [,32] [,33] [,34] [,35] [,36] [,37]
## id        20    21    22    23    24    25    26    27    28    29    30    31    32    33    34    35    36    37
## degree     4     4     4     5     5     5     4     3     4     3     4     4     5     5     4     3     3     3
##        [,38] [,39] [,40] [,41] [,42] [,43] [,44] [,45] [,46] [,47] [,48] [,49]
## id        38    39    40    41    42    43    44    45    46    47    48    49
## degree     4     6     5     4     4     4     4     3     3     5     4     6
```

```r
mean(degree(sw))
```

```
## [1] 4
```

```r
hist(degree(sw))
```

]

.pull-right[

![](presentation_files/figure-html/alldegree-1.png)

]

---

### Comparing Degree distribution:

---

### Shortest Path 
#### Small-world

Between 5 and 34

.pull-left[

```r
sp534=shortest_paths(sw,5,34)
print(sp534$vpath[[1]])
```

```
## + 4/49 vertices, from 8abde4e:
## [1]  5  4 41 34
```

```r
E(sw,path=sp534$vpath[[1]])$color="blue"
E(sw,path=sp534$vpath[[1]])$width=2
plot(sw, vertex.size=6, layout=layout_in_circle)
```
]

.pull-right[
![](presentation_files/figure-html/shortpathSW2-1.png)
]

---

### Shortest Path 
#### Small-world

Between 1 and 12

.pull-left[

```r
sp112=shortest_paths(sw,1,12)
print(sp112$vpath[[1]])
```

```
## + 4/49 vertices, from 8abde4e:
## [1]  1  2 44 12
```

```r
E(sw,path=sp112$vpath[[1]])$color="red"
E(sw,path=sp112$vpath[[1]])$width=2
plot(sw, vertex.size=6, layout=layout_in_circle)
```
]

--
.pull-right[
![](presentation_files/figure-html/shortpathSW-1.png)
]

---

### Shortest Path 
#### Full graph

Between node 3 and 45

.pull-left[

```r
sp345=shortest_paths(fg,3,45)
print(sp345$vpath[[1]])
```

```
## + 2/50 vertices, from 9ef0377:
## [1]  3 45
```

```r
E(fg,path=sp345$vpath[[1]])$color="blue"
E(fg,path=sp345$vpath[[1]])$width=4
plot(fg, vertex.label=NA, layout=lfg)
```
]

.pull-right[
![](presentation_files/figure-html/shortespathFG2-1.png)
]

---

### Shortest Path

#### Small-world

What if we compute all shortest path?

---

### Shortest path

#### Scale-free graphs

Between 4 and 48

.pull-left[

```r
E(ba)$color="grey"
E(ba)$width=1
sp448=shortest_paths(ba,4,48)
print(sp448$vpath[[1]])
```

```
## + 4/50 vertices, from f42cac5:
## [1]  4  1  3 48
```

```r
E(ba,path=sp448$vpath[[1]])$color="red"
E(ba,path=sp448$vpath[[1]])$width=2
plot(ba, vertex.size=degree(ba), vertex.label=NA, layout=baln)
```
]

.pull-right[
![](presentation_files/figure-html/scalefree1-1.png) 
]

---

### Shortest path
#### Scale-free graphs

Between 10 and 34

.pull-left[

```r
sp1034=shortest_paths(ba,10,34)
print(sp1034$vpath[[1]])
```

```
## + 5/50 vertices, from f42cac5:
## [1] 10  7  1  4 34
```

```r
E(ba,path=sp1034$vpath[[1]])$color="blue"
E(ba,path=sp1034$vpath[[1]])$witdh=2
plot(ba, vertex.size=degree(ba), vertex.label=NA, layout=baln)
```

```r
E(ba)$color="grey"
E(ba)$width=1
```
]

.pull-right[
![](presentation_files/figure-html/scalefree2-1.png) 
]

---

### Shortest path
#### Scale-free graphs

---

### Comparing Shortest Paths

---

### Betweenness centrality

How many shortest paths pass through each node?

.pull-left[

```r
graphs=list(sw=sw,scalefree=ba,random=rgg)
allbetween=lapply(graphs,betweenness)
aldst=range(sapply(allbetween,function(gr)range(density(gr,from=0)$y)))

plot(1,1,type="n",ylim=aldst,xlim=range(unlist(allbetween)*1.1),main="compare betweenness centrality",xlab="centrality",ylab="")
for(gr in seq_along(graphs))
    lines(density(betweenness(graphs[[gr]]),from=0),lwd=3,col=gr)

legend("topright",fill=adjustcolor(1:3,.4),legend=names(graphs))
```
]

.pull-right[
![](presentation_files/figure-html/compBetwen-1.png) 
]

---

### Centrality measureS

- Degree : which node has most in our out links

- Betweenness : which mode has most SP going through it...

- Page Rank : Google Algorithm

- And mode ; how are they distributed? how 'centralised' is the network...

---

Comparing degree & betweenness centrality

---

### Clustering Coefficient, modularity and communities
#### Clustering coefficient:

.pull-left[

```r
# Creating a simple undirected graph with edge colors
g <- graph.formula(A - B, C - D, D - A, B - D)

# Assign colors to edges
E(g)$color <- "red"
E(g)$width <- 3
glayout=layout_in_circle(g)

plot(g, edge.color=E(g)$color,layout=glayout)
```
]

.pull-right[
![](presentation_files/figure-html/clustering-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Clustering coefficient:

.pull-left[

```r
g_complete <- make_full_graph(vcount(g))

plot(g_complete, layout=glayout)

plot(g, edge.color=E(g)$color,layout=glayout,add=T)
```
]

.pull-right[
![](presentation_files/figure-html/clusteringComplete-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Clustering coefficient:

.pull-left[

```r

# Calculate the global clustering coefficient
global_cc <- transitivity(g, type="global")

plot(g_complete, layout=glayout,main=paste0("clustering coeff (global):",global_cc))
plot(g, edge.color=E(g)$color,layout=glayout,add=T)
```
]

.pull-right[
![](presentation_files/figure-html/clusteringCompleteGlo-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Clustering coefficient:

.pull-left[

```r

# Calculate the global clustering coefficient
local_cc <- transitivity(g, type="local")

par(xpd=T)
plot(g_complete, vertex.label="",vertex.size=0,layout=glayout,main=paste0("clustering coeff (global):",global_cc))
plot(g, edge.color=E(g)$color,layout=glayout,add=T,vertex.size=10)
text(x=glayout[,1],y=glayout[,2],labels=paste("local cc=",round(local_cc,digit=2)),pos=c(4,3,2,1))
```
]

.pull-right[
![](presentation_files/figure-html/clusteringCompleteLoca-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Modularity:

.pull-left[

```r
g <- make_graph(edges=c(1,2, 2,3, 3,1, 4,5, 5,6, 6,4, 1,4, 4,7, 5,7, 6,7 ), directed=FALSE)
glay=layout_nicely(g)
plot(g,layout=glay)
```
]
.pull-right[
![](presentation_files/figure-html/clus1-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Modularity:

.pull-left[

```r
membership <- c(1,1,1,2,2,2,2)
md=modularity(g,membership)
plot(g,vertex.color=membership,layout=glay,main=paste("modularity:",round(md,digit=2)))
```
]
.pull-right[
![](presentation_files/figure-html/clus2-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Modularity:

.pull-left[

```r
membership <- c(1,2,1,2,1,2,1)
md=modularity(g,membership)
plot(g,vertex.color=membership,layout=glay,main=paste("modularity:",round(md,digit=2)))
```
]

.pull-right[
![](presentation_files/figure-html/clus3-1.png)
]

---

### Clustering Coefficient, modularity and communities
#### Community detection

```r
allgraphs=list(sw,ba,rgg)
alllayouts=list(layout_in_circle,baln,rggln)
for(i in seq_along(allgraphs)){
	gr=allgraphs[[i]]
	lay=alllayouts[[i]]
	clgr=cluster_louvain(gr)
	gr$palette=categorical_pal(8)
	V(gr)$color=clgr$membership
	plot(gr,layout=lay,main=paste0("clustering:",round(transitivity(gr),digit=2),", modularity:",round(modularity(clgr),digit=2)))
}
```

---
class: inverse,middle,center

# Moving back to real world

## The Zachary's Karate Club

---

### The Zachary's Karate Club

Links between 34 members of a karate club. Given the club's small size, each club member knew
everyone else. Sociologist Wayne Zachary documented 79 pairwise links  between members who regularly interacted outside the club.

.pull-left[
![zacharymatrix](images/zacharymatrice.png)
]
--
.pull-right[
![zacharymatrix](images/zacharynetwork.png)
]

---

### The Zachary's Karate Club

.pull-left[

Conflict and split:

- A conflict between the club’s president and the instructor split the club into two.
- About half of the members followed the instructor ad the other half the president, a breakup that
unveiled the ground truth, representing club's underlying community structure.
- Today community finding algorithms are often tested based on their ability to infer these two communities from the structure of the network before the split.
]
--
.pull-right[
![zacharymatrix](images/zacharyoutput.png)
]

---

### The club of Zachary's karate club

.pull-left[
<div style="height: 100%;width:250px"> 
<blockquote class="twitter-tweet" " >This is the icing on the cake! I received the Zachary Karate Club Club trophy 🏆 for my talk in <a href="https://twitter.com/MscxNetworks?ref_src=twsrc%5Etfw">@MscxNetworks</a>! Thank you <a href="https://twitter.com/PiratePeel?ref_src=twsrc%5Etfw">@PiratePeel</a> for carrying it with you to Salina! <a href="https://t.co/153e3LBU9W">pic.twitter.com/153e3LBU9W</a>&mdash; Clara GM (@claragranell) <a href="https://twitter.com/claragranell/status/1038022679545176064?ref_src=twsrc%5Etfw">September 7, 2018</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script> 
</div>
]
.pull-right[
<div style="height: 100%;width:250px">

<blockquote class="twitter-tweet">Today I had the opportunity to meet the 2023 winner of the prestigious Zachary Karate Club Award. Congratulations to <a href="https://twitter.com/l_gajo?ref_src=twsrc%5Etfw">@l_gajo</a> for the impressive achievement! Well deserved <a href="https://twitter.com/hashtag/netsci2023?src=hash&amp;ref_src=twsrc%5Etfw">#netsci2023</a> <a href="https://twitter.com/netsci2023?ref_src=twsrc%5Etfw">@netsci2023</a> <a href="https://t.co/Vo5Fon3PjB">pic.twitter.com/Vo5Fon3PjB</a>&mdash; Leonardo Rizzo (@LnrdRizzo) <a href="https://twitter.com/LnrdRizzo/status/1678407578504626176?ref_src=twsrc%5Etfw">July 10, 2023</a></blockquote> <script async src="https://platform.twitter.com/widgets.js" charset="utf-8"></script> 
</div>
]

---

### Visualisation

.pull-left[

```r
zach <- make_graph("Zachary")
plot(zach,vertex.size=30,layout=lay,xlim=range(lay[,1]),ylim=range(lay[,2]),rescale=F)
```
]

.pull-right[

![](presentation_files/figure-html/zachary1-1.png)

]

---

### Who are  the 'important' nodes?

Degree of the nodes

.pull-left[

```r
degree(zach)
```

```
##  [1] 16  9 10  6  3  4  4  4  5  2  3  1  2  5  2  2  2  2  2  3  2  2  2  5  3  3  2  4  3  4  4  6 12 17
```

]

.pull-right[

![](presentation_files/figure-html/plotdegree-1.png)

]

[//]: # (```{r})
[//]: # (ggraph(zach, 'igraph', layout=lay,circular = F) + )
[//]: # (geom_edge_link() +)
[//]: # (coord_fixed() + )
[//]: # (geom_node_point( color = 'steelblue', size = degree(zach)) +)
[//]: # (ggforce::theme_no_axes())
[//]: # (```)

---

### Who are  the 'important' nodes?

betweenness centrality: number of shortest paths through a vertex

.pull-left[

```r
btw=betweenness(zach)
print(btw)
```

```
##  [1] 231.0714286  28.4785714  75.8507937   6.2880952   0.3333333  15.8333333  15.8333333   0.0000000  29.5293651
## [10]   0.4476190   0.3333333   0.0000000   0.0000000  24.2158730   0.0000000   0.0000000   0.0000000   0.0000000
## [19]   0.0000000  17.1468254   0.0000000   0.0000000   0.0000000   9.3000000   1.1666667   2.0277778   0.0000000
## [28]  11.7920635   0.9476190   1.5428571   7.6095238  73.0095238  76.6904762 160.5515873
```

```r
btw=betweenness(zach)
V(zach)$color=20+btw/3
zach$palette=rev(colorRampPalette(c("red","blue"))(max(btw/3+20)))
plot(zach,vertex.size=20+btw/3,layout=lay,xlim=range(lay[,1]),ylim=range(lay[,2]),rescale=F,vertex.label=NA)
```

]

.pull-right[

![](presentation_files/figure-html/plotbetweeness-1.png)

]

---

### Degree Distribution

```r
hist(dgre, breaks=1:vcount(zach)-1, main="Histogram of node degree: Zach")
hist(btw,breaks=20 , main="Histogram of node degree: Zach")
```

---

### Community Detection

- Community detection based on edge betweenness: High-betweenness edges are removed sequentially (recalculating at each step) and the best partitioning of the network is selected.

- Community detection based on modularity : split my group until finding split that maximise moudlarity

```r
 ceb <- cluster_edge_betweenness(zach)
 clo <- cluster_louvain(zach)
plot(ceb,zach,layout=lay)
plot(clo,zach,layout=lay)
```

---

### Visualize results:

.pull-left[
![](presentation_files/figure-html/unnamed-chunk-23-1.png)
]
--
.pull-left[
![zachresult](images/zacharyoutput.png)
]

---
class: inverse,middle,center

# And in Archaeology?

---

### Example and application

.small-leftm[

From Brughmans & Peeples 2023

]

.small-leftm[

- What can be our nodes?
{{content}}

]

- What can be our edges?
{{content}}

- What hypotheses can we explore?

- Which metrics to represent them?

.small-leftm[

(From Amati et al. 2020)

]

---

### Example and application

#### Geographical aspect: Visibility maps/distance/least cost path....

(From Brughmans & Brandes 2017)

.center[
<iframe src="https://www.google.com/maps/embed?pb=!1m18!1m12!1m3!1d3277.648135892445!2d-61.05046062492987!3d16.32169315631034!2m3!1f0!2f0!3f0!3m2!1i1024!2i768!4f13.1!3m3!1m2!1s0x8c6cc6b4feafb7c7%3A0x60d7aa9da94e6af0!2sMorne%20Souffleur!5e1!3m2!1sen!2suk!4v1709127838637!5m2!1sen!2suk" width="300" height="225" style="border:0;" allowfullscreen="" loading="lazy" referrerpolicy="no-referrer-when-downgrade"></iframe>
]

---

### Example and application

#### Geographical aspect: Visibility maps/distance/least cost path....

(From Brughmans & Brandes 2017)

#### Production/consumption:

---

### Example and application

Between sites similarity :

Counting artefacts between site, which artefact do they share or not? Dissimilarity/Similiarity matrix.

Metal Artefact: 
.center[ <img src="images/metartefact.png" width="460"> ]

.small-leftm[
<img src="images/metalnet.png" width="220">
]

--
.small-leftm[
<img src="images/metalcomu.png" width="220">
]

.small-leftm[
- decoration
- type of burial
- presence absence
- Brainerd Robinson and other index (Simpson diversity,....)
{{content}}

]

--
- [https://github.com/simoncarrignon/bronze-age-ssr](github.com/simoncarrignon/bronze-age-ssr)

---

### Problems & Limitation

Easy tool/Great visualisation/Simplify complex dataset & problems

---

### Problems & Limitation

- Prignano et al. 2017: similarity metrics

- Daems et al. 2023: time averaging

- ...

---

### Time Averaging

---

### Incomplete Networks

.small-left[
<a href="https://commons.wikimedia.org/wiki/File:Internet_map_1024.jpg#/media/File:Internet_map_1024.jpg"><img src="https://upload.wikimedia.org/wikipedia/commons/d/d2/Internet_map_1024.jpg" alt="Internet map 1024.jpg" width="280"></a> by Matt Britt using data from the OPTE project
]

.small-left[

&nbsp; Real, but complete network 
 {{content}}
]

&nbsp; In archaeology?

.small-left[
![](presentation_files/figure-html/smallworld-1.png)
]

.small-left[
![](presentation_files/figure-html/scalefree-1.png)
]

---

<img src="presentation_files/figure-html/unnamed-chunk-24-1.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-2.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-3.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-4.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-5.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-6.png" width="20%" /><img src="presentation_files/figure-html/unnamed-chunk-24-7.png" width="20%" />
---

---
class:middle,center
<img src="presentation_files/figure-html/unnamed-chunk-25-1.png" width="40%" /><img src="presentation_files/figure-html/unnamed-chunk-25-2.png" width="40%" />

---

---
class: inverse,center,middle

# Take Home Message

---
class: inverse,center,middle

## Networks are _nice_

## Networks are _easy_

## Networks are _powerful_

## Networks are (were?) _hype_

#### Networks are _nice_: _easy_ to use & _powerful_, but don't let the (any) _hype_ fool you