I cannot
believe, that this is our last blog … Or maybe last blog for this class and who
knows, I can possibly continue blogging on my own, now that I feel some comfort
in blogging! Picking the last topic to blog was a bit of a challenge, as you
wanted it to possibly be the best you have written! After some debating, I
decided that blogging about web analytics would be a pretty good choice – as it
was said in our class it is the most exciting aspect of a data analytics
course! We know that in today’s tech world we pretty much do everything online
which means we leave tracks that can later be analyzed – we can see how we are
connected on LinkedIn and Facebook, what pages we like, what groups we join,
what we are interested in based on our posts and tweets, etc.
First,
let’s familiarize my readers with what is actually a network. A network is a
social structure composed of entities, or vertices, and the relationships, or
edges, among the entities. Understanding networks actually helps us understand
the world which is a complex system of people and things that are interacting
among themselves or each other. How? In more technical terms, networks helps us
understand various phenomenon like relationship formation among customers,
information diffusion and disease propagation!
Network
analysis is an emerging Business Intelligence technique that is being more and
more used in social network analytics, telecommunication analytics, criminal
intelligence, banking, human resources, and many more industries.
A
network is visually represented with nodes (vertices) and edges (links), as
shown on the picture to the right. The nodes in the network represent the
entities in a system, while the edges represent the relationships between the
nodes. Nodes can be people, universities, companies, webpages, animals, plants,
Facebook pages, etc. On the other hand, edges are the “labels of a
relationship” and they can be either directed or undirected as well as weighted
or unweighted.
It
is worth mentioning that there are many types of networks, but the primary
types are single and two mode networks. In a single mode network there is only
one kind of node, while in a two mode network, you can have more than one node,
or specifically two kinds of nodes.
Networks
like anything else, have properties that help us understand their structure.
Some of the main structural properties of networks include centrality measures,
density, clustering coefficient, cliques, etc.
In this blog we will discuss some not so complex ideas of network
analysis that have significant consequences.
The
main idea when dealing with networks is the connection. Differences among
individuals or populations in how connected they are can be significant in
understanding their behaviour and attributes. More connections mean that
individuals or populations are exposed to more information, and probably more
diverse information, which means that they can be more influential and more
influenced by others.
The
next approach deals with the distance between individuals and/or populations. As majority of individuals are not
usually directly connected to most other individuals in a population, sometimes
it is necessary to study connections beyond immediate connections, of course in
addition to the density of direct connections. For example, an
individual may be able to reach out to most other members of a population with
almost no effort by sharing the message with his/her friends, who then tell
their friends, and so on until "everyone" knows. Others might have
more difficulty spreading the message, as the people they share it with might
not be well connected, thus the message would not go far. This notion also
applies to populations that differ in how close individuals are to others and
those differences can help us understand diffusion (the spreading of something more widely),
homogeneity (the quality or
state of being homogeneous), solidarity (unity or agreement of feeling or action, especially among
individuals with a common interest; mutual support within a group), etc.
Finally,
I want to end this blog with a listing of few social network analysis softwares
found on KD Nuggets site:
●
Gephi is an interactive visualization and exploration platform for all kinds of networks and
complex systems, dynamic and hierarchical graphs. Runs on Windows, Linux and
Mac OS X. Gephi is open-source and free.
●
Centrifuge offers analysts and investigators an
integrated suite of capabilities that can help them rapidly understand and
glean insight from new data sources, visualize discoveries by interacting with
data, collaborate to draw conclusions.
●
R is a general purpose analytics tool, but several
libraries are available for social network analysis. These include degreenet,
RSeina, PAFit, igraph, sna network, tnet, ergm, Bergm, hergm, latentnet and
networksis. Each provides specialised functionality and for people familiar
with R represent a rich set of resources.
●
SocNetV (Social Networks Visualizer) is a
cross-platform, user-friendly tool for the analysis and visualization of Social
Networks. It lets you construct networks (mathematical graphs) on a virtual
canvas, or load networks of various formats (GraphML, GraphViz, Adjacency,
Pajek, UCINET, etc). Also, SocNetV enables you to modify the social networks,
analyse their social and mathematical properties and apply visualization
layouts.
References
Hanneman,
R. A., & Riddle , M. (n.d.). Introduction to Social Network Methods.
Retrieved from ucr.edu:
http://faculty.ucr.edu/~hanneman/nettext/C7_Connection.html
Top 30 Social
Network Analysis and Visualization Tools. (2015, June). Retrieved from
KDnuggets:
http://www.kdnuggets.com/2015/06/top-30-social-network-analysis-visualization-tools.html/3
