Computational Social Science

Computational social science is an interdisciplinary field that leverages mathematical algorithms, advanced data analysis, and computational modeling to study and predict human behavior and social dynamics. By integrating techniques from computer science, statistics, and social sciences, computational social science offers powerful insights into complex social phenomena, enabling researchers to analyze large-scale data and simulate interactions in real time.

The Basic Idea

Imagine that you are a medical researcher, interested in predicting epidemic breakouts to prevent diseases from spreading. If you wait for an outbreak to be confirmed—meaning that an individual has received a positive test—then it’ll be very likely that the disease will have already started to make its way across the population. 

What if there was a way to gauge who has the disease without having to wait for a positive test result? Well, there is a way—through computational social science!

Computational social science leverages computational methods and big data to study and predict human behavior in society. By applying tools from computer science to analyze social patterns, this approach helps us decode the complexities of human interactions. With our widespread use of online apps and social media platforms, researchers and analysts now have access to vast amounts of data that can be used to answer important social science questions. 

Let’s return to our example. A computational social scientist can estimate and predict epidemics using Google searches. Google is often the first place we turn when we have symptoms, long before we take a test to confirm. Computational social science would map Google trends with officially documented outbreaks to predict cases that have not yet been confirmed and to anticipate where an outbreak may occur. By comparing the Google searches in areas with confirmed cases to similar Google searches happening in other regions, this method can help accurately predict disease spread.¹

In this example, computational social science filled the gap of confirmed outbreaks through other available data, as not everyone will get a test when they feel sick. This approach enables us to make links between data and human behavior to understand complex social phenomena. As technology continues to evolve through advancements in machine learning and AI, computational social science can also help us understand how such algorithms are shaping our behavior and society.² 

Social Science: The study of people and their behavior within society.

Computational Science: A field that applies computer science and software engineering principles to solve scientific problems.⁴

Data Science: A multidisciplinary field that uses complex analytical and statistical methods and machine learning algorithms to analyze large datasets.

Big Data: A large collection of data in terms of volume, velocity, and variety. 

Agent-Based Model: Computer simulations used to examine the relationship between people, things, places, and time. Computer scientists create digital representations of people, known as agents, which are assigned attributes and programmed to interact with other agents in the environment to study the effects.⁵ 

The Tipping Point: The critical point in a situation, process, or system, at which an idea or behavior spreads and snowballs and leads to an unstoppable effect or change. Computational social science often seeks to find what the tipping point is to hypothesize when an unstoppable effect on human behavior would take place.⁶  

Social science was born during the Age of Enlightenment, a movement in the 17th and 18th centuries where people began using reason to explain human behavior and the world around them.⁷ The rise of capitalism inspired people to understand humans in the context of the society in which they lived, rather than seeing behavior as a product of an individual, isolated and separate from the environment in which they operate.

Realizing how complex adaptive systems of society are led to the advent of agent-based modeling in the late 20th century, which simulates how groups of agents interact with each other within an environment. In 1970, John Conway pioneered agent-based modeling through his computer simulation called the “Game of Life,” where the interaction of cells (representing humans) updated its state based on that of its neighbors, mimicking how people interact in societies. This was one of the first instances of using computer simulations to understand human behavior. What was unique about Conway’s Game of Life was that it was a zero-player game, meaning that no human intervention was required for effects to take place, demonstrating that by nature, individual agents will affect one another.⁸ 

In recent years, thanks to further development of data science and computational technology, social science has experienced a paradigm shift, leading to the creation of a new branch, computational social science. The term was popularized by computer scientist David Lazer and his colleagues in 2009, who understood the vast potential of computational social science for large-scale social analysis in the age of big data. Our widespread use of the internet and social media has led to increased access to billions and billions of pieces of information that can be analyzed to understand and predict human behavior.

“Our movements in public places may be captured by video cameras, and our medical records stored as digital files. We may post blog entries accessible to anyone, or maintain friendships through online social networks. Each of these transactions leaves digital traces that can be compiled into comprehensive pictures of both individual and group behavior, with the potential to transform our understanding of our lives, organizations and societies.” – Lazer et al. in their 2009 paper, “Computational Social Science”⁹

John Conway

English mathematician well-known for his 1970 “Game of Life” model which replicated human behavior within a society. This simulation helped to demonstrate that individuals usually affect and are affected by other agents around them. Conway is well known for his contributions to number theory, game theory, and coding theory.⁸

David Lazer

American professor of computer science and political science, Lazer is often credited with popularizing the term computational social science in 2009. Now that researchers had access to vast amounts of data providing digital records of human behavior, Lazer believed this could be combined with advances in computer memory and processing power, allowing analysis that could generate new insights for social sciences.⁹

Thomas Schelling

Economist and Nobel laureate well-known for his contributions to nuclear strategy. Schelling also created a model to help explain urban racial segregation in his 1971 Article, “Dynamic Models of Segregation,” in which he proposed the concept of the “tipping point.” This is the moment where an idea or behavior crosses a threshold, “tips,” and spreads. In his application of urban racial segregation, the tipping point was the moment at which a mixed neighborhood would switch to a segregated one. The tipping point model became known as the “Schelling Model of Segregation” and a quintessential example of computational social science.⁷ 

Harrison White

White began his career as a physicist and later turned to sociology to study how social networks shape individual’s everyday lives. White conceptualized humans as nodes within social networks, allowing him to apply mathematical models to social structures to explain individual and societal behavior.¹⁰ 

Computational social science has the potential to transform the way that we gather and analyze data, expanding the way we understand human behavior. We live in complex social systems that are difficult to comprehend—and previously, research on human behavior within a society relied on mainly self-reported data, limiting our ability to research human interactions at a large scale.11 As stated by Cornell social scientist Michael Macy, computational social science has “been really transformative. We were limited before to surveys, which are retrospective, and lab experiments, which are almost always done on small numbers.” ¹¹

By applying computational models and algorithms to collect and examine data, computational social science draws patterns that were previously difficult to observe, such as anticipating an epidemic by analyzing Google search trends. As computational social science can draw on vast amounts of big data at a quick pace, it can provide us with real-time analysis that informs us about both micro and macro structures of society and the agents within them. 

Using simulations, computational social scientists can explore the impact of interventions, assisting policymakers in making more informed decisions. Computational social science has been pivotal in areas where it is difficult to collect reliable data. For example, developing countries often lack sophisticated data collection strategies, making it hard to identify population characteristics such as poverty. Through computational science, we can use mobile phone metadata, such as call frequency and duration, to predict the wealth and poverty levels. This strategy hinges on the assumption that wealthier people tend to make more calls to a wider range of locations compared to poorer people due to phone call costs. In this way, computational science provides an inexpensive method to assess economic conditions and give insight into where interventions are required. 

As is common with any field that relies heavily on behavioral data, there are several ethical concerns with computational social science. The very reason why computational social science is able to transform the way we analyze human behavior, through the collection and analysis of big data, is also the very reason this technique compromises privacy. Without gaining informed consent and providing transparency, people remain unaware of how their data is being used or even who has access to it. The field has developed so rapidly that it has outpaced regulatory frameworks, making it unclear who owns the data, who should be allowed to use it, and for what purposes. Without these frameworks in place, there is increased risk of data misuse. 

Additionally, computational social science uses big data to fill in the gaps of recorded data—meaning sometimes, it tries to draw conclusive patterns that do not actually exist. Can we confidently claim that the frequency and duration at which someone uses their mobile phone is indicative of their level of wealth? Although the correlation might be there, the causation can not be determined. In the same vein, while big data allows a vast quantity of information, that does not necessarily guarantee quality data. Yet, studies have demonstrated that computational social science models have the potential to make predictions that match real-life records, giving it a degree of credibility. 

While policies and guidelines need to catch up with the field of computational social science, there is no disputing the profound positive impacts that the field can have in understanding social phenomena, helping drive evidence-based decision-making for addressing social issues. 

Infectious Ideas

According to the mere exposure effect, people are more likely to adopt a new idea the more they are exposed to it. However, computational social science has shown that this theory isn’t 100% accurate. 

In 2012, John Kleinberg, a computer scientist, collected data from around 900 million Facebook users to study how ideas are spread within social networks, as well as what makes someone likely to join a social media platform.11 What Kleinberg found is that people’s decision to create a Facebook account was not based on how many friends they had already using the platform, but with how many distinct groups those friends were a part of. For example, if someone had one friend from work, one from university, and one family member on Facebook, they were more likely to join than someone who had three friends from the same group on the platform. 

While understanding why someone joins a social media network or not may seem like the most important application of computational social science, it demonstrates how ideas spread, and can be applied to political races, public-health campaigns, and marketing.

Combating Misinformation

During the COVID-19 pandemic, you were sure to land on information about the pandemic everytime you turned on your TV, opened your laptop, or scrolled on your phone. There was an unprecedented amount of research effort by the global scientific community, making it difficult to sort through the explosion of studies exploring the diverse aspects of the pandemic. 

With communication dispersing so quickly online, there was an increased risk of misinformation being spread, countering efforts to enforce effective countermeasures against the virus. Many online platforms were even publishing information that was not yet reviewed or confirmed.

In order to try and prevent the spread of misinformation, medical researchers developed a software framework based on the collection of analysis of 140,000 research articles about the COVID-19 pandemic, categorizing them into six topics: Molecular Pharmacology and Biomarkers, Molecular Virology, Epidemiology, Healthcare, Clinical Medicine, and Clinical Imaging. The analysis helped to categorize the information and applied a scoring system to identify the top peer-reviewed articles that were most relevant and trustworthy.

After the initial analysis of over 14,000 research articles, the software framework was able to automatically gather new articles from multiple sources, as well as information about the source of publication and authors, in order to predict the chances that the article would pass a peer-review process. The use of computational social science allowed researchers to sort COVID-19 articles by the likelihood that it shared relevant, trustworthy information that would pass a peer-review, helping governments and policy makers to make timely decisions about the most effective restriction measures to lessen the spread of the virus.

Government Nudging in the Age of Big Data

One of the advantages of computational social science is that it allows us to see how interventions will impact a population through trial and error, before an intervention or nudge is applied to the real world. In this article, our writer Johnny Hugill, explores how data science, machine learning, and predictive analytics have led to smarter policy implementations in Britain. 

Why tech products should be designed alongside psychologists

Technology is now deeply embedded into our society—it can change the way we feel, view the world, and even the choices we make. In this article, our writer Juan Manuel Contreras argues that because technology has a profound impact on our behavior, technological companies should collaborate with psychologists to better understand the effect their innovations will have on people. 

1. Jarynowski, A., Wójta-Kempa, M., Płatek, D., & Czopek, K. (2020). Attempt to understand public-health relevant social dimensions of COVID-19 outbreak in Poland. BMC Public Health, 20(1), 1144. https://doi.org/10.1186/s12889-020-10106-8
2. UBC Centre for Computational Social Science. (n.d.). What is computational social science? Retrieved from https://ccss.arts.ubc.ca/about/what-is-computational-social-science/
3. Social Science Insights. (December 2015). Social science quote: Computational social sciences. Retrieved July 9, 2024, from https://socialscienceinsights.com/2015/12/31/social-science-quote-computational-social-sciences/
4. Rouse, Margaret. (August 2017.). Computational science.Techopedia. Retrieved July 9, 2024, from https://www.techopedia.com/definition/6579/computational-science
5. Columbia University Mailman School of Public Health. (n.d.). Agent-based modeling. Retrieved July 9, 2024, from https://www.publichealth.columbia.edu/research/population-health-methods/agent-based-modeling
6. Flache, A., & Macy, M. W. (2017). From the Margins to the Mainstream: Simulating the Evolution of Norms Governing Indirect Reciprocity. Journal of Artificial Societies and Social Simulation, 20(3), Article 15. Retrieved July 9, 2024, from https://www.jasss.org/20/3/15/15.pdf
7. Liberto, Daniel. (May 2024). Social Science: What It Is and the 5 Major Branches. Investopedia. https://www.investopedia.com/terms/s/social-science.asp#toc-history-of-social-science
8. Simudyne. (n.d.). Conway's Game of Life and the birth of agent-based modeling. Retrieved July 9, 2024, from https://www.simudyne.com/resources/conways-game-of-life-and-the-birth-of-agent-based-modeling
9. Lazer, D., Pentland, A., Adamic, L., Aral, S., Barabási, A.-L., Brewer, D., Christakis, N., Contractor, N., Fowler, J., Gutmann, M., Jebara, T., King, G., Macy, M., Roy, D., & Van Alstyne, M. (2009). Computational social science. Science, 323(5915), 721-723. https://doi.org/10.1126/science.1167742
10. Rosenwald, Michael (June 2024). Harrison White, Groundbreaking (and Inscrutable) Sociologist, Dies at 94. The New York Times. https://www.nytimes.com/2024/06/12/books/harrison-white-died.html
11. Giles, Jim. (2012). Computational social science: Making the links. Nature, 488, pg. 448-450. https://doi.org/10.1038/488448a
12. Napolitano, F., Xu, X., & Gao, X. (2022). Impact of computational approaches in the fight against COVID-19: An AI-guided review of 17,000 studies. Briefings in Bioinformatics, 23(1), bbab456. https://doi.org/10.1093/bib/bbab456