JohnI am finally done with the 12th class of my MBA degree, which was Introduction to Social Computing. In this class, we went through a lot of topics, but a few of my favorite are networks, introduction of data analytics using R and Python programming languages, and my favorite of the class – game theory. I wanted to give a few snippets of information I found interesting from the class.
Networks
The network aspect of the class started real broad and we basically reviewed relationships between nodes. For instance, we learned about strong and weak ties and how they work with triadic closure, clustering coefficient, and neighborhood overlap.
Strong and Weak Ties
The idea here is basic, there can be strong and weak ties in a network. However, the strength of a weak tie can be overlooked for its strength. For instance, weak ties can often produce the best job leads via an acquaintance rather than a close friend. Another example of how a weak tie can be beneficial is to be a valuable conduit to hard-to-reach parts of a network, such as a celebrity.
Triadic Closure
Triadic closure is the idea that if there are two people in a social network that have a common friend, there is a likelihood they will become friends in the future. This is usually due to opportunity, trust, and incentive. In the image above, there is a high likelihood of node B and C becoming friends because they both are friends with node A.
Clustering Coefficient
Clustering coefficient is way to determine the probability of a node having two randomly selected friends that are friends with each other. The range is 0 to 1, where 0 means none are friends and 1 means all nodes are friends. The way to determine this is the following steps:
- Determine number of friends node A has
- Find possible pairs of friends among node A’s friends using n(n-1) / 2; where n = number of nodes connected to A (from step 1)
- Find number of pairs that are actually friends
- Determine node A’s clustering coefficient with the following equation:
- (number of pairs that are actually friends) / (possible pairs of friends among node A’s friends)
- or (step 3 value) / (step 2 value)
In this scenario, step one would be 3, step two would be 3, step three would be 1, therefore, number four/clustering coefficient for node A would be 1/3.
Neighborhood Overlap
In a neighborhood overlap, a person is trying to find the ratio of nodes that are neighbors with both A and B divided by the nodes that are neighbors of either A or B:
number of nodes who are neighbors of both A and B / distinct number of nodes who are neighbors with either A or B
In the scenario above, the neighborhood overlap of A and B would be 1/3.
Data Analytics w/ R and Python
In the data analytics portion of the class, we were able to utilize R and Python programming languages to derive information from different sources such as Twitter, YouTube, and Yelp data. We used R Studio and Spyder via Anaconda for Python.
An example of finding correlation between twitter friends and followers based on searching on aapi:
Game Theory
Game theory is the study of human behavior in strategic settings. In essense, the payoff of a person’s decision also depends on the choice of other people’s decision. Examples of this is a soccer penalty kick and tennis. Decision of the goalie can alter the payoff of the shooter. In tennis, one player’s decision to hit to a certain area can affect if and how the other reacts. A better way to explain this is with a payout matrix.
Payout Matrix
In a payout matrix, there are decisions made by stakeholders and each decision has a payoff. In the matrix below, the Shooter can decide to either shoot left or right, and the Goalie can decide to jump left or right – from their perspectives. The payoff values on the left in each cell are the payoffs for the Shooter and the values on the right belong to the Goalie. In this scenario, if the shooter shoots right and the goalie jumps left, the Shooter has a 0 payoff because the Goalie will have a 100 payoff; meaning the Goalie will save/block the shot. In either decision, it is best for the Goalie to dive the way the Shooter is shooting. I understand it’s pretty obvious and outrageous payoffs, but it’s an over-simplistic example.
While the class was only seven weeks, there was a lot of beneficial information that changed my way of thinking. I found myself traveling down rabbit holes in trying to apply this to my current work and has me wanting to start other analytical projects. Though intrigued, my 13th class starts soon, so I’ll be honest and say I won’t really have the time. Having said that, only three more classes to go!
Thanks for reading!