Carnegie Mellon University

36-200: Reasoning with Data

In today’s world, an understanding of Data Science is critical for success in the Humanities and Social Sciences.

In Reasoning with Data (36-200), students explore key questions like:

The world is filled with uncertainty, constant flux, and variation. Moreover, we are all prone to conceptual biases and perceptual illusions. How can we say we know anything about the world amidst this uncertainty, variation, and bias? In Reasoning with Data, we begin to explore methods and concepts in making knowledge that can address and overcome these obstacles.

In Reasoning with Data, students make essential connections to history asking questions like “why was Napoleon instrumental in linear regression?” or “how was the London plague fundamental to modern data analysis?”

In Reasoning with Data, students investigate, model, describe, and explain how nations compare across an array of socioeconomic and demographic measures.

In Reasoning with Data, students have the opportunity to interact with written publications in unique ways, exploring topics like:

  • High-dimensional word analysis of presidential speeches
  • Range of writing level of Wikipedia
  • Evidence for multiple authors of the Bible
  • Quantitative evidence for multiple pseudonyms of Samuel Clemens (Mark Twain)
  • Connection between Gulliver’s Travels and the origin of statistics

How can we detect, describe, and measure bias and discrimination like ‘redlining’ and other restrictive practices? How can we precisely measure social class differences, in contexts ranging from modern-day New York to the Titanic disaster? We learn how in Reasoning with Data.

In Reasoning with Data, we explore cases that reveal bias like:

  • Do people perform better when given positive reinforcement or negative reinforcement (and does the answer depend on the task being performed)?
  • After faced with a failure, precisely how do optimistic people tend to react versus pessimistic people?
  • What effect does expectation and framing have on outcome bias?
  • Do women or people of color perform worse on an exam if told beforehand that people like them tend to perform poorly?

Social creatures group themselves into complex networks. How do we begin to understand and model such complex multidimensional social networks, and what are the limitations of such models?  Students in Reasoning with Data explore, describe, and discuss such networks and their assumptions and limitations.

In Reasoning with Data, students work with a wide array of real datasets, from contexts as diverse as social media to global studies to fruit flies. The data are real, complex, and authentic; and students are empowered with both the tools and the freedom to deeply explore questions of immediate relevance to them and to the larger world.  

For instance, in a recent semester, some actual questions that students have developed and explored (completely on their own) in full-length data analysis projects in the course include:

  • Racial inequity in online dating
  • Historical connections between social media and the Black Lives Matter movement
  • Sexual orientation and depression