Skip to content

Welcome

Reading Image

IPHS 290: Computational Cultural Analytics
Prof Elkins & Chun
Integrated Program for Humane Studies
Kenyon College
Fall 2022

Overview

Computational cultural analytics refers to the collection, analysis, modeling and visualization of data for the exploration of contemporary and historical cultures. Driven by domain expertise from study in traditional fields like political science or literary criticism, computational methods augment research by incorporating the analysis of data created by constituent entities and their interactions (e.g. social media, government documents, contextualizing financial narratives, etc).

This is an advanced methods course focused on learning to work with complex data types. The topics learned in this course will dramatically expand the possibilities of your future research in cultural analysis and generalize to analysis in any domain. This course presumes some coding experience or the introductory course to Digital Humanities, IPHS 200 Programming Humanity.

The most popular stereotype of data is in the form of relatively small and structured numeric data commonly found in spreadsheets. In fact, the the most common and fastest growing type of data is unstructured data: raw text, sound, images and videos. We compile, transform, model, visualize and analyze both of these datatype in IPHS 300: AI for the Humanities.

This course introduces methods to work with different and popular composite datatypes including geospatial, time series and network graphs. These big 3 composite data types each have distinct internal structure based upon spatial, temporal or semantic correlations. Each of these data types have their own distinctive methods, libraries and critiques which we explore via both leading academic DH scholarship and practical technical implementations.

Here is an outline of the course:

  • We begin the course with visualizing and analyzing geospatial data via maps and charts. We’ll build our skills with web scraping and API’s to create original datasets from websites and services like Twitter, Reddit and Instagram.

  • Then we’ll study a variety of natural language processing (NLP) subjects from data wrangling to vectorization and topic modeling to state-of-the-art (SOTA) deep neural nets.

  • Narrative sentiment analysis and SentimentArcs provide a NLP framework to explore the particularities of time-series analysis.

  • Graph theory and network analysis introduce new methods to visualize, connect and compute metrics of related entities based on network representations. We study hands-on projects like analyzing the social network of Game of Thrones and trying to classify who’s tweeting: Trump or Trudeau.

  • The final week will focus on a variety of options to share presentations, ML/AI models or applications on the web. This includes several free and paid web hosting options from no-code static blog platforms to dynamic web full-stack virtual servers on cloud hosts like Amazon Web Services.

  • Students produce a wide variety of individualized stand-alone projects to demonstrate specific learned skills. The course culminates in a final class project based on each student's intellectual interests and personal passions. These compelling narratives are reconceived and implemented using any combination of the technologies and techniques presented throughout the semester.