363-1091-00L  Social Data Science

Semester Spring Semester 2017
Lecturers D. Garcia Becerra
Periodicity yearly course
Language of instruction English

Abstract Social Data Science is introduced as a set of techniques to analyze human behavior and social interaction through digital traces.
The course focuses both on the fundamentals and applications of Data Science in the Social Sciences, including technologies for data retrieval, processing, and analysis with the aim to derive insights that are interpretable from a wider theoretical perspective.
Objective A successful participant of this course is able to
- understand a wide variety of techniques to retrieve digital trace data from online data sources
- store, process, and summarize online data for quantitative analysis
- perform statistical analyses to test hypotheses, derive insights, and formulate predictions
- implement streamlined software that integrates data retrieval, processing, statistical analysis, and visualization
- interpret the results of data analysis with respect to theoretical and testable principles of human behavior
- understand the limitations of observational data analysis with respect to data volume, statistical power, and external validity
Content Social Data Science (SDS) provides a broad approach to the quantitative analysis of human behavior through digital trace data.
SDS integrates the implementation of data retrieval and processing, the application of statistical analysis methods, and the interpretation of results to derive insights of human behavior at high resolutions and large scales.
The motivation of SDS stems from theories in the Social Sciences, which are addressed with respect to societal phenomena and formulated as principles that can be tested against empirical data.
Data retrieval in SDS is performed in an automated manner, accessing online databases and programming interfaces that capture the digital traces of human behavior.
Data processing is computerized with calibrated methods that quantify human behavior, for example constructing social networks or measuring emotional expression.
These quantities are used in statistical analyses to both test hypotheses and explore new aspects on human behavior.

The course is structured in three main parts. First, collective behavior is analyzed with respect to time trends, distributions, and information sharing. The second part focuses on the processing and analysis of text, applying and validating sentiment analysis methods. The third part covers empirical social network analysis based on online social network data, covering both topological and dynamic aspects of social networks.

The course will cover various examples of the application of SDS:
- Search trends to measure information seeking
- Popularity signals and social influence
- Microblogging data to measure mood
- Digital markets and cryptocurrencies
- Sentiment analysis across various online media
- Twitter network analysis

The lectures include theoretical foundations of the application of digital trace data in the Social Sciences, as well as practical examples of data retrieval, processing, and analysis cases in the R statistical language from a literate programming perspective. Weekly exercise classes provide practical skills and discuss the solutions to exercises that build on the concepts and methods presented in the previous lectures.
Lecture notes The lecture slides will be available on the Moodle platform, for registered students only.
Literature See handouts. Specific literature is provided for download, for registered students only.
Prerequisites / Notice Participants of the course should have some basic background in statistics and programming, and an interest to learn about human behavior from a quantitative perspective.

Prior knowledge of R, information retrieval, or information systems is not necessary.

Self-study tasks are provided as home work and build on technical and theoretical content explained in the lectures.

Weekly exercises (45 min) are used to discuss the solutions and guide the student.