Introduction to Computational Social Science

Summary

In this video, Chris Bale introduces Computational Social Science (CSS), an interdisciplinary field combining computational techniques with social science theories to analyze large-scale data from sources like social media and digital archives. He discusses how digital trace data, such as search patterns and social media activity, helps track human behavior, exemplified during the COVID-19 pandemic. CSS also addresses new questions, like misinformation and political polarization. Despite challenges like data access and ethical concerns, the field offers exciting potential for addressing global issues, from public health to climate change, while evolving as a multi-disciplinary area of study.

Takeaways

• 😀 CSS is an interdisciplinary field that applies computational techniques to large datasets to advance theories of human behavior.
• 😀 The explosion of data over the past 20 years, especially from social media, smartphones, and digitized archives, provides new insights into human behavior.
• 😀 Digital trace data (e.g., Google search data) allows researchers to study human behavior in real-time, such as tracking the spread of pandemics like COVID-19.
• 😀 While digital trace data holds great potential, it is not perfect and can be influenced by external factors, such as advertising algorithms that skew health-related searches (e.g., Google Flu).
• 😀 Despite its flaws, combining digital trace data with other data sources, such as surveys or socioeconomic data, can reveal meaningful patterns and insights.
• 😀 One key challenge in CSS is data access. Much of the data is controlled by corporations (e.g., Facebook, YouTube) and is not freely available for research purposes.
• 😀 There is a tension between prediction and explanation in CSS. While data science focuses on prediction, social science often aims to explain why behaviors occur.
• 😀 CSS is in its early stages, and its definition will evolve over time. Researchers are still defining the goals and scope of the field.
• 😀 New questions arising from digital behaviors, such as misinformation on social media and its impact on politics, are at the forefront of CSS research.
• 😀 CSS's potential extends beyond existing data to answer new questions about societal issues, including climate change, public health, and political polarization.
• 😀 CSS is a dynamic, multi-disciplinary field, combining insights from computer science, sociology, psychology, economics, and other disciplines to address complex global challenges.

Q & A

What is computational social science (CSS)?

Computational social science is an interdisciplinary field that applies computational techniques to large data sets from sources like social media, smartphones, and historical archives to study human behavior and society.

How has the COVID-19 pandemic influenced computational social science?

The pandemic has highlighted both the strengths and weaknesses of CSS. It showcased the ability of CSS to track human behavior in real-time using digital trace data (e.g., Google search trends), but also revealed challenges, such as data misinterpretation and the limitations of passive data collection.

What is digital trace data, and why is it important for CSS?

Digital trace data refers to information generated through online activities, such as search queries or social media interactions. It is important for CSS because it provides a rich, real-time record of human behavior, which was previously difficult to capture with traditional methods like surveys.

Why did Google Flu fail, and what can we learn from it in CSS?

Google Flu failed because its model misinterpreted data due to biases introduced by Google's own advertising algorithms. This highlights the importance of carefully interpreting digital trace data and considering external factors that may influence online behavior.

What role does socioeconomic status play in CSS research, particularly during the COVID-19 pandemic?

Socioeconomic status plays a significant role in behavior during crises like the COVID-19 pandemic. For example, wealthier individuals were more likely to shelter in place because they had the resources to work from home or access larger living spaces, demonstrating the intersection of social factors with behavioral data.

What are some of the primary challenges faced by computational social science researchers?

One of the main challenges in CSS is data access. A large portion of valuable data is stored by corporations (e.g., Facebook, YouTube) and is not publicly accessible for research. Additionally, there is a tension between using CSS for prediction versus explanation of human behavior.

How has the field of CSS evolved over time?

CSS has evolved through three stages: (1) early agent-based modeling in the 1980s, (2) using computational power to analyze massive datasets from digital sources, and (3) integrating theory from social sciences to generate generalizable insights about human behavior.

What is the significance of combining different data sources in CSS?

Combining different data sources, like digital trace data with surveys or administrative records, can provide richer, more accurate insights into human behavior. For example, combining census data with location data during the pandemic helped researchers understand how social factors like income influenced people's ability to social distance.

What new questions is CSS now poised to address?

CSS is now addressing questions related to new digital phenomena, such as the spread of misinformation on social media, the impact of digital trace data on public health, and how new technologies like social media shape human behavior in politics and society.

How does CSS contribute to the development of theories of human behavior?

CSS helps develop theories of human behavior by using computational tools to analyze large-scale data, testing existing social science theories, and generating new insights. This can lead to more generalizable or semi-generalizable knowledge about human behavior, especially in the context of digital technologies.