What Does a Computational Social Scientist Do? Notes From a Computational Social Scientist

Computational Social Science is a discipline that brings together three core components: Computer Science, Social Science, and Insights.

Originally posted on Medium.


Computational Social Science is a discipline that brings together three core components:

  • Computer Science — the use of computer tools to handle huge amounts of data.
  • Social Science — the study of human and societal behavior using the scientific method.
  • Insights — the generation of knowledge using analytical and storytelling techniques.

Whenever someone asks what I do, more often than not, I find people looking back at me in confusion. But I recently got to share with my company what my graduate program is all about, so I figured I would elaborate on these notes for anyone interested in understanding what a “Computational Social Scientist” does.

I have been a Computational Social Scientist for years — sometimes unknowingly — and I frequently use “Social Data Scientist” as a synonym. Although they are technically different disciplines, I’d summarize both in the same sentence:

“The use of computers and social science principles to make sense of the world.”

The sentence can be broken down into three elements: Computer Science, Social Science, and Making Sense of the World.

Computer Science

Think of Computer Science as the tools a Computational Social Scientist will use. Even though (arguably) all careers in this modern age employ computers on a daily basis, Computational Social Science puts it front and center.

The whole premise of integrating computers into a social scientist’s workflow is that we have huge amounts of online data (i.e., “big data”), and computers offer a way of handling it. For example, where before someone would manually analyze a handful of surveys or interviews, now we can use software to automate and analyze millions of them.

It is worth noting that a Computational Social Scientist does not necessarily know how to build hardware or software, but it does imply you know how to use them in your favor — e.g., you may not know how to build a website from scratch but you probably know how to programmatically extract and analyze information from it.

To handle data, you often use programming languages — such as R, SQL, or Python. If unfamiliar with these terms, it can help to imagine that, just as you may speak and write English to communicate with others and get stuff done, you use programming languages to communicate with your computer and deal with information.

One last note is that computers are not only a tool but also the subject of study—more on that in the next section.

Social Science

Think of Social Science as the subject of study for a Computational Social Scientist. In other words, a CSS studies human and societal behavior, which often leads to a CSS being a subject matter expert in a particular social science domain, such as economy, politics, or psychology.

This is also where the “science” of it all comes in. Being a scientist involves creating knowledge following the scientific method — in natural sciences, this may be easier because you can accurately measure the liters of water in a container, but in social science, how do you measure the level of sympathy among a group of people?

Just because something is hard to measure, it doesn’t mean we should stop studying it. Since social science data (and real-life) can be messy, we rely on the scientific method to develop methodologically sound experiments and models. In other words, you become highly methodic and critical about approaching a problem.

In a similar vein, this helps to be critical of other people’s work, so if you are reading a research article or analyzing a dataset, you should be able to question whether its findings stand to reason or are borderline pseudo-science.

Lastly, a convergence of Computer Science and Social Science can be studying topics such as the ethics of artificial intelligence or algorithmic biases. In these instances, you are essentially taking one of the tools you work with and analyzing it from a social science perspective.

Making Sense of the World

Think of Making Sense of the World as generating knowledge or insights. To me, this is probably the most relevant aspect because it represents the “so what?” of CSS. What will you find once you have the tools (computers) and the subject of study (social science topic)?

It is also the more practical part of Computational Social Science, as this allows you to extract meaningful insights and create impactful stories. In this regard, data analytics and storytelling go hand-in-hand with generating insights.

Although debatable, this is where I’d say statistics come into play. Once you have a social science dataset and know how to work your way around computers, you can conduct statistical analysis to come to conclusions. For example, if you have a survey with demographic data, you can use a machine learning algorithm to divide — or even predict — people’s income.

But perhaps most importantly, generating insights entails activities such as creating reports, dashboards, slide decks, and posters. It is the culmination of your work as you can visually represent your conclusions in a condensed fashion.

Written and graphical deliverables are key elements that your peers, supervisors, and stakeholders will use to judge your work and possibly make decisions — e.g., giving research grants or implementing business recommendations.

In Conclusion

Computational Social Science is a discipline that uses computer tools to study a social science topic and generate insights. You use computers and programming languages to deal with big data; you use the scientific method to study a social science domain; and you use analytics and storytelling techniques to come to conclusions.

This is evidently a practical and oversimplistic approach toward Computational Social Science. I barely scratched the surface on areas such as statistical modeling or public opinion — subjects that smarter people than me can spend a lifetime understanding.

Furthermore, Computational Social Science is a relatively new discipline. With emergent technologies such as artificial intelligence, we can only expect it to evolve according to our society’s needs — but after all, that’s the whole point of this field.

I hope these insights add some value to your journey. Any additional comments are more than welcome, and I will continue to update this article — or create follow-ups — as I dive deeper into Computational Social Science.

Miguel Curiel
Miguel Curiel
Data | Insights | Growth

Bridging the gap between social science and data science. Passionate about people, insights, and stories.