Cloudlight

• A library for analysing complex networks.

• Based in industry-wide tools such as Python and NetworkX.

• Includes experimental features such as new network analysis parameters and privacy attack simulations

This work was financed by CONICET (Argentinian Council of Research), in the context of a phd grant awarded to José Orlicki under supervision of José Ignacio Alvarez-Hamelin and Pablo Fierens (Beca Tipo I CONICET, Convocatoria 2009).

Documentation

• Tutorial - Getting started
• Some Graph samples

What is new about Cloudlight?

The original features of Cloudlight are:

• Extension of basic functionality of NetworkX (that works on RAM) to work with indexed graphs on a SQL persistant database.

• Graphs with weighted edges over persistant SQL.

• Rudimentary graph compression.
  

• Finite computation of internal scaling dimension and connectivity dimensionof graphs based on concepts of fractal dimensión on complex graphos (paper).  

• Statistical analysis of triangles, average neighbor degree, eccentricity and average path length on persistant or volatile complex graphs.

• Statistical analysis of node neighborhood size for persistant complex grapohs.

• Implementation of complex graph graph k-coreness shell-index over SQL.

• Automated analysis of all statistical graph parameters included in the library and 2D plotting. Including degree, clustering index, average neighbor degree, eccentricity, average path length, internal scaling, connectivity and k-corenessshell-index .

• Implementation of heuristic simulation for privacy attacks on social networks systems. Based on bibliography heuristics and centered on scenarios where the node degree is a-priori visible(paper).

•  Implementation of heuristic simulation for privacy attacks on social networks systems, centered on scenarios where the node degree is a-priori not visible. 

• Set of scripts to build SQL databases for graphs, analyze them, run simulations and dump numerical results.