This is an old revision of the document!

Software released by Online Social Networks project

  • Clique Estimation: two Python scripts that demonstrate the estimators described in our paper Estimating Clique Composition and Size Distributions from Sampled Network Data. The first script implements two types of unbiased estimators of clique size distributions, one of which exploits labeling of sampled nodes neighbors and one of which does not require this information. Additionally, it supports the compositions of cliques by node attributes (only supports binary node attributes, such as gender). The second script demonstrates how to prepare the data for input to the first script. More specifically, it receives as input a known graph, sampling parameters (sampling method, sampling size, replacement type), and clique distribution preferences (labeling, attributes). It then appropriately samples egonets from the given graph and calculates the maximal clique distribution for each sampled egonet.
  • 2.5K-Graphs: two software packages to demonstrate the algorithms and estimators described in our paper 2.5K-Graphs: from Sampling to Generation. The first package receives as input a random walk graph sample and estimates the degree-dependent clustering coefficient distribution and network average clustering coefficient. The second package implements all the algorithms and estimators within classes “Estimation” and “Generation”. It receives as an input a fully known graph and then simulates a random walk graph sample of given size. The class “Estimation” provides functions that estimate the degree-dependent clustering coefficient (CCK) and joint degree distribution (JDD). The class “Generation” provides functions that generate a 2.5K graph given specific CCK and JDD distributions.
  • Geosocialmap: a web-based tool that visualizes geo-social data. More information can be found in our paper Coarse-Grained Topology Estimation via Graph Sampling and the M.Sc. thesis GeoSocialMap Visualization
  • Graph sampling: A set of functions to sample nodes of a graph with replacements (Simple Random Walk, Weighted Random Walk, Metropolis Hastings Random Walk, Uniform Independent Sampling, Weighted Independent Sampling) and corresponding estimators.
  • Facebook Applications: protype crawlers of the Facebook user profiles in 2008 and user coverage simulator. More information can be found in our paper Poking Facebook: Characterization of OSN Applications