In my previous blog I discussed what sociophysics is and why advanced methods from can be applied to social systems. In this blog, I will give a few typical examples of sociophysics' applications to social systems.
(1) Opinion propagation is a very popular topic in sociophysics. The opinions can be about specific brands, products, services, or political preferences. Much as in condensed-matter physics, where particles can propagate a signal (i.e., information) through interactions with each other, in social systems the connected individuals can propagate opinions by sharing them with each other through word of mouth. What is important: As a common starting point, sociophysics assume a mechanism at work in how individuals share opinions (e.g., a person convinces his or her neighbor in a social network). Then the opinions propagate through the social network following this mechanism, and the propagation dynamics are analyzed. Thus, under some conditions a minority can become a majority. One of the important results from the dynamics is a “final" distribution of opinions that can be stable until new opinions and mechanisms will prevail. Such stable states can have many different forms, e.g., two main opinions are shared by the whole society, or one becomes dominant.
(2) Self-organized criticality (e.g., catastrophic event) is an extremely powerful theory applicable to any process of sufficient duration. In physics, the key idea is that some tensions are accumulated and then find resolution through a critical (catastrophic) event, just to clear the ground for new accumulation. For example, adding new sand grains to the tip of a sand pile will eventually result (after one more grain) in a sudden avalanche, after which a new, flatter pile will start forming, eventually building to a new avalanche. Processes like that are familiar in economics and business, e.g., a housing bubble.
(3) Theory of networks starts from graph theory and presents a new picture of the world, where everything is in a mutual connectivity of different types. Perhaps the best-known example is the “Six Degrees of Separation" theory, which states that everyone is six (maximum) or fewer steps away, by introduction (a friend of a friend), from any other person in the world. Another popular example is powerful neural networks that are considered as learning systems and used as a form of Machine Learning algorithms. Traditional statistics often oversimplifies or misses complex connectivity in the social world and, instead, substitutes correlations or distances between objects.
(4) Simulation models are associated with the growing popularity of agent-based and system dynamics models, which have their roots in physics. In such simulation models physics concepts are used as first principles to build a detailed social model (with all important real-life factors included), which is then run (i.e., it lives as in a real life). During the simulation, a modeler is analyzing the corresponding social dynamics and considering convergence processes and “final" (quasi-equilibrium) states of social systems. Calibration of a simulation model, i.e., fitting of parameters, allows the establishment of a solid correspondence of modeling outputs with observable real-life parameters, like sales. From there, what-if analysis and objective-based optimization conclude the process.
(5) Mediaphysics is a methodology to model population mind-set distributions in a mind-set metric space between choices (brands, products, etc.). It can adopt almost any important real-life effects like word of mouth, advertising, economic factors, opinion propagation with its inertia, long- and short-term memory, entropic disorganizing factors, etc. Processes of mediaphysics can be simulated in terms of system dynamics of simulation modeling. Mediaphysics uses advantages of both statistics and sociophysics with organic integrity. I plan to write a special blog dedicated to this methodology, which we developed a few years ago.
Sociophysics can definitely provide significant enhancements to business analytics in addition to traditional statistics. This is my second blog in a series on sociophysics. In the next blog I will discuss some of the most important differences between sociophysics and traditional statistics.