Social NetWrok Analysis

Backend implementation of social relations and clusters with basic visualiser to mine a given social network, detect influenciers followers and community topic

• Implement graph based data analysis function to mine a given social network
• Some analysis e.g. Centrality, LanceWilliams HAC, Hierarchical Clustering, Agglomerative Clustering, Closeness/Betweenness Centrality

Input Data

from a text file and build a directed weighted graph using the graph ADT, samples are shown in /graphs for example 0, 4, 5 represents 0->4 with a weight 5

Stage 1 build graphs

Implemented adjavency List representation 

Stage 2 Dijkstra's algorithm

In discovering "influencers", ew need to repeatedly find shortest paths between all pairs of node.
By Implementing Dijkstra's algorithm to discover shortest paths from a given source to all other nodes in the graph based on Priority Queue 

Stage 3 Centrality Measures for Social Network Analysis

Centrality measures 
higher betweenness measure often correspond to influencers

Degree Centrality

defined as the number of links incident upon a node

Closeness Centrality

Betweenness Centrality

Stage 4

How to use the testing interface

To use Dijkstra get the shortest path output for your search on the input graph, and compare output files with correct solution in /dijkstarPaths

$ ./testDijstark [input file]   # for any input file
$ ./testDijstra.sh 1    # test search on graph1

To use Centrality

Run ${./testCentralityMeasures} to get HELP to use the test interface

Usage: ./testCentralityMeasures [file] [flag]

Flag	def
d	degree centrality
do	degree out centrality
di	degree in centrality
c	closeness centrality
b	betweenness centrality
bn	betweenness centrality normalised
v[m]	trigger Graph Vis with mode [m]

0    : DEFAULT
1    : DEGREE_IN
2    : DEGREE_OUT
3    : DEGREE
4    : CLOSENESS
5    : BETWEENNESS

To use graphVis interface

call graphVis(g, DEFAULT)

LanceWilliamsHAC analysis of clusters of users' social relationship

the sample visualization as shown below