300275 Soziologische Methoden - quantitativ: Methods and applications in Computational Social Science (S) (SoSe 2022)

Computational social science or computational sociology is an approach that combines computer science and social science to use new types and sources of data to answer sociological and socio-political questions. This seminar will provide an introduction to the basic concepts required when approaching computational social science from a social science background.

One half of the course will focus on useful technical background knowledge, models and algorithms applied in the field of computational social sciences (methods). This includes but is not limited to: Tool selection (programming languages, frameworks), dealing with Big Data, research data management, web scraping, API usage, computational speed and performance, algorithm complexity, basic machine learning algorithms, an introduction to Deep Learning and finally text data processing and analysis.

The second half of the course covers applications, i.e., empirical work belonging to the field of computational social sciences is discussed, with important data types such as online and social media data. Students will develop their own research question and project, which they will work on during the semester (possibly in groups with several students). A special focus will be put on the possibility of visualising the research results.
Above all, we will also discuss which questions can be answered with the help of Computational Social Science and its methods, where possible limitations lie and what constitutes good Computational Social Science research. Students will use the statistical software R to apply their newly acquired knowledge in practice. However, we will also briefly discuss the choice of other tools, e.g. Python (because of some specific frameworks) and Julia (to implement algorithms that are not available in other software). In general, students can use any tool or programming language, but most examples will be demonstrated in R. Students will get access to the computing cluster of the Faculty of Sociology making it possible to get some valuable experience in analysing big data in a high performance computing environment.

The seminar will be held in English. The course credit (Studienleistung) will also be in English. Examination (Prüfungsleistung/Hausarbeit i.e term paper) can be completed in English or German. Practical knowledge in handling data is helpful (e.g., knowledge of univariate statistics or previous use of other statistical software such as SPSS or Stata). Knowledge of R is an advantage, but a very brief introduction will be given to show students with no prior knowledge some resources to catch up on their own.

Please note that this is a 'major seminar' (4 hours per week), which means that you will receive credit for a full module if you complete the assignments and the term paper (normally two assignments in two different seminars and a term paper in one of the two seminars are required). This means that Statistical Sciences students can complete the module '30-SW-ESo Empirische Sozialforschung' by attending this course only, given they complete the assignments and the term paper. Sociology Master students can receive credit for a complete module '30-M-Soz-M3c Soziologische Methoden' in the same way.

The preliminary syllabus is as follows:

05.04.2022: Opening session (short tour of the topics covered, assignments, term paper, course structure and tools) & possible applications of computational social science

12.04.2022: Introduction to R

19.04.2022: Collecting online data (HTML, Webscraping, APIs [Twitter and SDMX]) & potentials and problems of collecting online data (total error framework, ethics and privacy)

26.04.2022: Proecessing text data and data visualization

03.05.2022: Research data management & open and reproducible science

10.05.2022: Tools in computational social science & discussing your research questions

17.05.2022: Spatial data & linking data sources (regional data, official statistics or survey data)

24.05.2022: Supervised machine learning and examples (Predicting energy consumption)

31.05.2022: Unsupervised machine learning and examples (TBA)

14.06.2022: Deep learning and examples (TBA)

21.06.2022: Topic modeling (Blei et al. 2003, Roberts et al. 2014) & examples (OECD priorities in education)

28.06.2022: Word2vec & examples (Gender stereotypes over time)

05.07.2022: Big data, computational speed and performance, alrogithmic complexity (big-oh notation) & coding and Q/A session

12.07.2022: Closing session & group project