In this section, users are able upload network files in either txt, or csv format.
Network files should be formatted as follow; Node 1, Node 2 and weight.
Each column may be separated with: comma (,) semicolon (;) or tab character (\t)
followed by the end of line (\n) character at the end of each line.
Files that successfully uploaded to the server can be found in workspace tab menu.
Please, enable cookies on your web browser in order to proceed.
Users are able to upload either sample data, by pressing �Upload Sample Data� button (1), or upload their networks by pressing �Choose file� button (2), selecting multiple files, and then pressing �Upload to Workspace� button (3).When data are successfully uploaded, they can be found at (4). By pressing �Empty Workspace� button (5), uploaded data are removed from the server. Users are also able to download and visualize their uploaded data (6).
Users are able to navigate to the graph either by holding left mouse button and scrolling mouse or by using the navigation panel (1). By selecting a node in the graph, 1st neighborhood connections are only visible (2). Please be aware that by clicking outside of the graph, the whole network is revealed again, thus navigation buttons should be preferred. Initial network is in random layout, however users are able to select between random, circular or Fruchterman-Reingold layout (3). Nodes of the network can also be found and selected using �select node� (4). Node labels and weights can be interactively modified using control panel (5).
2. Feature Networks Caclucations
In this section, users are able to generate ten global network characteristics � features (1).
1. Degree of Distribution (mean-median-max): The histogram of nodes with the number of their adjacent edges
2. Clustering Distribution (mean-median-max): The histogram for the maximal connected component sizes
3. Closeness Centrality (mean-median-max): Measures how many steps are required to access every other node from a given node
4. Centralization: A graph level centralization measure from the centrality scores of the nodes
5. Centrality of Centralization: The histogram of the centrality scores of the nodes
6. Betweenness: The shortest path from one node to another (mean-median-max)
7. Normalised Betweenness: The normalised shortest path from one node to another (mean-median-max)
8. Degree of Undirected Assortativity: Whether two nodes tend to connect to each (handles network as undirected graphs)
9. Degree of Directed Assortativity: Whether two nodes tend to connect to each (handles network as directed graphs)
10. Alpha Centrality (mean-median-max): Calculates the alpha centrality of all vertices in the network
11. Jaccard similarity based on the nodes� or edges' indexes.
When pressing �Generate Features� (2), features are calculated for all available uploaded networks. If data are well generated, an appropriate message is generated in section �General Network Features� (3), were users are able to delete or download them in csv format. In addition, features can be downloaded (1) or viewed (2) at the Network Analysis Results section that is also generated in the same page.
3. Clustering Networks
In this section users are able to apply clustering based on the following parameters (1):They can chose: between hierarchical and affinity propagation clustering, among a variety of distances (Euclidean, Maximum, Manhattan, Canberra or Minkowski), among clustering linkage (Ward, Single, Complete, Average, Mcquitty, Median or Centroid), among desired outcome clusters, using only the selected by checkbox generated features.
After applying clustering, the following clustering image (dendrogram) is generated.
By clicking the Visualize button, all networks with their calculated distances can be displayed. The larger the connection of two networks, the more bold the interconnected edge will be displayed