Login to your profile!



No account? sign up!

Synthie     (Labeled Networks)

Download network data

This network dataset is in the category of Labeled Networks



Visualize Synthie's link structure and discover valuable insights using the interactive network data visualization and analytics platform. Compare with hundreds of other network data sets across many different categories and domains.

Metadata

CategoryLabeled Networks
CollectionLabeled Networks
Tags
ShortLabeled Networks
DescriptionThis network contains the following comma separated text files:
Synthie is a synthetic data sets consisting of 400 graphs. The data set is subdivided into four classes. Each node has a real-valued attribute vector of dimension 15 and no labels. We used the following procedure to generate the data set: First, we generated two Erdös-Rényi graphs using the G(n,p) model with p0.2 and n: 10. For each graph we generated a seed set S_i for i in {1,2 of 200 graphs by randomly adding or deleting 25% of the edges. From these seed sets we generated two classes C_1 and C_2 of 200 graphs each by randomly sampling 10 graphs from S_1 cup S_2 and randomly connecting these graphs. For C_1 we choose a seed graph with probability 0.8 from S_1 and with probability 0.2 from S_2. The class C_2 was generated the same way but with interchanged probabilities. Finally, we generated a set of real-valued vectors of dimension 15 subdivided into two classes A and B using the make_classification method Scikit-learn. We then subdivided C_i into two classes C^A_i and C^B_i by drawing a random attribute from A or B for each node. For class C^A_i, we drew an attribute from A if the node belonged to a seed graph of seed set S_1, and from B otherwise. Class C^B_i was created the same way but with interchanged seed sets.

Please cite the following if you use the data:

@inproceedings{nr,
     title={The Network Data Repository with Interactive Graph Analytics and Visualization},
     author={Ryan A. Rossi and Nesreen K. Ahmed},
     booktitle={AAAI},
     url={http://networkrepository.com},
     year={2015}
}

Note that if you transform/preprocess the data, please consider sharing the data by uploading it along with the details on the transformation and reference to any published materials using it.

Network Data Statistics

Nodes36.6K
Edges161.7K
Density0.000240981
Maximum degree48
Minimum degree2
Average degree8
Assortativity-0.171457
Number of triangles125.8K
Average number of triangles3
Maximum number of triangles68
Average clustering coefficient0.0801794
Fraction of closed triangles0.0615987
Maximum k-core11
Lower bound of Maximum Clique4

Network Data Preview

Interactive visualization of Synthie's graph structure

Interactively explore the networks graph structure!

  • Use mouse wheel to zoom in/out
  • Mouseover nodes to see their degree
  • Drag network to see more details

Loading...

Interactive Visualization of Node-level Properties and Statistics

Tools for Interactive Exploration of Node-level Statistics

Visualize and interactively explore Synthie and its important node-level statistics!

  • Each point represents a node (vertex) in the graph.
  • A subset of interesting nodes may be selected and their properties may be visualized across all node-level statistics. To select a subset of nodes, hold down the left mouse button while dragging the mouse in any direction until the nodes of interest are highlighted.This feature allows users to explore and analyze various subsets of nodes and their important interesting statistics and properties to gain insights into the graph data
  • Zoom in/out on the visualization you created at any point by using the buttons below on the left.
  • Once a subset of interesting nodes are selected, the user may further analyze by selecting and drilling down on any of the interesting properties using the left menu below.
  • We also have tools for interactively visualizing, comparing, and exploring the graph-level properties and statistics.
Note: You are not logged in!
Please login or join the community to leverage the many other tools and features available in our interactive graph analytics platform.

Interactive Visualization of Node-level Feature Distributions

Node-level Feature Distributions

degree distribution

Loading...

degree CDF

Loading...

degree CCDF

Loading...

kcore distribution

Loading...

kcore CDF

Loading...

kcore CCDF

Loading...

triangle distribution

Loading...

triangle CDF

Loading...

triangle CCDF

Loading...

All visualizations and analytics are interactive and flexible for exploratory analysis and data mining in real-time and include the following features:

  • Degree, k-core, triangles, and triangle-core distributions. We include plots for each of the fundamental graph features and counts of the number with a particular property (i.e., number of nodes that form k triangles or have degree k, etc.)
  • We also include the CDF and CCDF distributions for each graph in the collection.
  • All visualizations and plots are zoomable. One may zoom-in or out on the data visualization using scrolling.
  • Panning. Users may also click anywhere on the plot and move the mouse in any direction to pan.
  • Adjust scale and other application dependent-parameters. All interactive visualizations may adjust the scale which is particularly important in certain types of graph data that contain highly skewed graph properties (power-lawed graphs and/or networks) such as degree distribution.

Discuss and Share

Collaborate and contribute to the first interactive and community-oriented data repository!

Share key insights, awesome visualizations, or simply discuss advantages of data, any observed or known properties, challenges, problems, corrections, and any other helpful comments! Post and discuss recent published works that utilize this dataset (including your own). Any and all feedback is welcome and encouraged.