Images as Data for Social Science Research
An Introduction to Convolutional Neural Nets for Image Classification

An introduction to how object and facial recognition, and visual sentiment analysis, can be used to process images as data.

Nora Webb Williams (Author), Andreu Casas (Author), John D. Wilkerson (Author)

9781108816854, Cambridge University Press

Paperback / softback, published 13 August 2020

75 pages, 32 b/w illus.
22.5 x 15.1 x 0.5 cm, 0.14 kg

Images play a crucial role in shaping and reflecting political life. Digitization has vastly increased the presence of such images in daily life, creating valuable new research opportunities for social scientists. We show how recent innovations in computer vision methods can substantially lower the costs of using images as data. We introduce readers to the deep learning algorithms commonly used for object recognition, facial recognition, and visual sentiment analysis. We then provide guidance and specific instructions for scholars interested in using these methods in their own research.

1. Introduction
2. Prerequisites for computer vision methods and tutorials
3. Introduction to CNNs for social scientists
4. Overview of fine-tuning a CNN classifier for images
5. Political science working example: images related to a Black Lives Matter protest
6. The promise and limits of autotaggers
7. Application: fine-tuning an open source CNN
8. Legal and ethical concerns in using images as data
9. Conclusion
10. References.

Subject Areas: Data mining [UNF], Data capture & analysis [UNC], Digital lifestyle [UD], Politics & government [JP], Society & social sciences [J], Research methods: general [GPS]