Author and Supervisor Network
Abdul Samad
May 20, 2021
Service as a Freelancer
If you want to hire me as a freelancer for your assignment or projects please use the following links.
- For Fiverr https://www.fiverr.com/dataanalyst_85
- For Upwork https://www.upwork.com/freelancers/~017667e5ef4347dcb4
1 - Define your nodes
The network contains 40 nodes with one attribute namely "Name". The nodes represents the author, supervisor and comittee member. Latter in the visualization, the yellow color corresponds to author and comittee members, while, the blue color represents the supervisor.
## + 39/39 vertices, named, from 55d34ea:
## [1] Ercan Eren Gulsun Yay Fatma Dogruel
## [4] Aysu Insel Zafer Tunca Necati Mumcu
## [7] Suut Dogruel Nami Nomer Osman Kucukahmetoglu
## [10] Turan Yay Merih Paya Ensar Yilmaz
## [13] Huseyin Tastan Burc Ulengin Nedim Sualp
## [16] Nesrin Sungur Melike Bildirici Turkel Minibas
## [19] Izzetin Onder Kaya Ardic Emel Yurt
## [22] Emin Karaaslan Mahmut Tekce Emin Koksal
## [25] Erhan Aslanoglu Ester Ruben Feride Gonel
## [28] Murat Cokgezen Ahmet Guner Sayar Murat Donduran
## + ... omitted several vertices2 - How did you define the relation between them?
In the network visualization at the end, author have a direct relationship with comittee members and supervisor. As, our network is weighted directed network, the weight between author and supervisor is 2, while, the weight between author and comittee members is 1. In the Above visualization of network, dotted edge representing the relationship between author and comittee members, while, solid edge representing the relationship between author and supervisor. Moreover, red color shows the direct relationship between author and supervisor, while, black color shows the relationship between author and comittee members.
E(g) # Edges of the network## + 44/44 edges from 55d34ea (vertex names):
## [1] Ercan Eren ->Murat Donduran Gulsun Yay ->Asuman Oktayer
## [3] Gulsun Yay ->Ensar Yilmaz Fatma Dogruel ->Huseyin Tastan
## [5] Aysu Insel ->Fazil Kayikci Zafer Tunca ->Senem Sahin
## [7] Necati Mumcu ->Kasim Eren Aysu Insel ->Mesut Karakas
## [9] Suut Dogruel ->Tolga Aksoy Nami Nomer ->Tunc Durmaz
## [11] Osman Kucukahmetoglu->Zeynep Kaplan Turan Yay ->Ozge Kama
## [13] Fatma Dogruel ->Murat Donduran Gulsun Yay ->Murat Donduran
## [15] Merih Paya ->Asuman Oktayer Suut Dogruel ->Asuman Oktayer
## [17] Ensar Yilmaz ->Asuman Oktayer Huseyin Tastan ->Asuman Oktayer
## [19] Suut Dogruel ->Huseyin Tastan Aysu Insel ->Huseyin Tastan
## + ... omitted several edges3 - Is your network directed or undirected, why?
Yes, the network is directed. We have representing the direct relation between author and supervisor as well as between author and comittee members. It is due to the supervision of author, as author can have only one supervisor. Moreover, it is due to the author, he is deadling with both supervisor and comittee members. So, the relationship is directed between these.
is.directed(g) # is netowrk directed?## [1] TRUE4 - Is your network weighted? If so, how do you define your weights?
The network is weighted. As the author is the only actor who has contact with both supervisor and comittee members. To distinguish the relationship, weight is added to the links. The weight for contact to the supervisor is 2, while, the weight for contact with the comittee member is 1.
is.weighted(g) # is network weighted>## [1] TRUE5 - Which of the following calculation are meaningful in your network? Explain the reason.
- Degree, in-degree, out-degree, weighted degree, weighted in-degree, weighted out-degree
As our network contain the direct relationship between author and supervisor as well as between author and comittee members. In the directed nework indegree and outdegree plays an important role. However, meaningfulness of the above measures depend on the scenario. If we want to see who has the high number of relationships then degree is meaningful? if we want to see which author or comittee member have the highest relationships? then outdegree and weighted outdegree meaningful. Similarly, if we want to see which comittee member have been spotted most of the time in meatings? then indegree is meaningful, instead of weighted indegree or other calculations. Because, every comittee member has the same level or weight.
- Which nodes have the highest values for them, why?
Top node by degree is Huseyin Tastan, by indegree is Asuman Oktayer, by outdegree is Gulsun Yay. Similarly, the top node by weighted degree is Huseyin Tastan, by weighted indegree is Asuman Oktayer and weighted outdegree is Gulsun Yay. First 15 scores of nodes for all the above calculations are presented below.
| Degree | Out_Degree | In_Degree | W_Degree | W_In_Degree | W_Out_Degree | |
|---|---|---|---|---|---|---|
| Ercan Eren | 1 | 1 | 0 | 2 | 0 | 2 |
| Gulsun Yay | 4 | 4 | 0 | 6 | 0 | 6 |
| Fatma Dogruel | 3 | 3 | 0 | 4 | 0 | 4 |
| Aysu Insel | 3 | 3 | 0 | 5 | 0 | 5 |
| Zafer Tunca | 1 | 1 | 0 | 2 | 0 | 2 |
| Necati Mumcu | 1 | 1 | 0 | 2 | 0 | 2 |
| Suut Dogruel | 3 | 3 | 0 | 4 | 0 | 4 |
| Nami Nomer | 1 | 1 | 0 | 2 | 0 | 2 |
| Osman Kucukahmetoglu | 1 | 1 | 0 | 2 | 0 | 2 |
| Turan Yay | 1 | 1 | 0 | 2 | 0 | 2 |
| Merih Paya | 1 | 1 | 0 | 1 | 0 | 1 |
| Ensar Yilmaz | 4 | 3 | 1 | 5 | 2 | 3 |
| Huseyin Tastan | 6 | 3 | 3 | 7 | 4 | 3 |
| Burc Ulengin | 1 | 1 | 0 | 1 | 0 | 1 |
| Nedim Sualp | 2 | 2 | 0 | 2 | 0 | 2 |
6 - Which of the following centrality measures are applicable to your network? Explain the reason.
- Degree, Betweenness, Eigenvalue, Closeness
Again it depends on the scenario in which you are going to find the central node. For our network, Degree, Betweenness and Closeness are applicable. If we look at the results, Huseyin Tastan has the highest degree and betweenness score. It means that Huseyin Tastan has the high number of relationships in the network, as well as he is the central node that is acts as a bridge between interaction of other nodes. Besides, he is close to all other nodes. However, Eigen centrality did not perforemd well on this network, as it 0 assigned all the nodes. Therefore, it is not applicable to our network.
- Which nodes have the highest values for them, why?
Top node by degree is Huseyin Tastan, by betweeness is Huseyin Tastan, by closeness is Aysu Insel. However, Eigen centrality could not produce better results by giving 0 to all the nodes. First 15 scores of nodes for all the centralities are presented below.
| Degree | Betweenness | Closeness | Eigen | |
|---|---|---|---|---|
| Ercan Eren | 1 | 0.0 | 0.0006920 | 0 |
| Gulsun Yay | 4 | 0.0 | 0.0007937 | 0 |
| Fatma Dogruel | 3 | 0.0 | 0.0007710 | 0 |
| Aysu Insel | 3 | 0.0 | 0.0007943 | 0 |
| Zafer Tunca | 1 | 0.0 | 0.0006920 | 0 |
| Necati Mumcu | 1 | 0.0 | 0.0006920 | 0 |
| Suut Dogruel | 3 | 0.0 | 0.0007508 | 0 |
| Nami Nomer | 1 | 0.0 | 0.0006920 | 0 |
| Osman Kucukahmetoglu | 1 | 0.0 | 0.0006920 | 0 |
| Turan Yay | 1 | 0.0 | 0.0006920 | 0 |
| Merih Paya | 1 | 0.0 | 0.0006925 | 0 |
| Ensar Yilmaz | 4 | 2.0 | 0.0007310 | 0 |
| Huseyin Tastan | 6 | 6.5 | 0.0007310 | 0 |
| Burc Ulengin | 1 | 0.0 | 0.0006925 | 0 |
| Nedim Sualp | 2 | 0.0 | 0.0007112 | 0 |
7 - Put the screen-shot(s) of your network that you think that gives the best look (be sure that you colored your picture and arrange the size of the nodes according to a meaningful measure and give the measure)
nodes <- data.frame(id = V(g)$name, title = V(g)$name, group = V(g)$type, font.size =rep(16,39))
edges <- get.data.frame(g, what="edges")[1:2]
vis.nodes <- nodes
vis.links <- edges
vis.nodes$shape <- "dot"
vis.nodes$shadow <- TRUE # Nodes will drop shadow
vis.nodes$title <- V(g)$name # Text on click
vis.nodes$label <- V(g)$name # Node label
vis.nodes$size <- degree(g)+25 # Node size
vis.nodes$borderWidth <- 2 # Node border width
vis.links$width <- 3 # line width
vis.links$font.size <- 15
vis.links$label <- as.vector(ifelse(E(g)$weight==1,1,2))
vis.links$dashes <- as.vector(ifelse(E(g)$weight==1,TRUE,FALSE))
vis.links$color <- palette()[E(g)$weight]
vis.links$arrows <- "to" # arrows: 'from', 'to', or 'middle'
vis.links$smooth <- FALSE # should the edges be curved?
vis.links$shadow <- FALSE # edge shadow
vis.nodes$color.border <- "black"
vis.nodes$color.highlight.background <- "orange"
vis.nodes$color.highlight.border <- "darkred"
k <- visNetwork(vis.nodes, vis.links, width = "100%", height = 700)
k <- visGroups(k, groupname = "Top", shape = "square",
color = list(background = "gray", border="black"))
k <- visGroups(k, groupname = "Simple", shape = "dot",
color = list(background = "tomato", border="black"))
kvisSave(k, file = "net.html", background = "white")Community Detection
wk <- walktrap.community(g)
plot(wk,g)