Big Data and Machine Learning in Sociology

The digital revolution and the resulting datafication of society have changed empirical social science research fundamentally. Enormous amounts of data can now be easily stored, managed and analyzed. Furthermore, the digital innovations of recent years allow the collection of data in various formats that were previously difficult to collect (e.g., georeferenced data, tracking or process data, intensive longitudinal data, social media text data). Finally, the increase in computational power as well as the maturation of software environments promoted the development of algorithmic solutions for complex statistical problems and prepared the way for the nascent field of computational social sciences at the intersection of social science research (sociology, economics, psychology, political science, criminology, communication studies etc.), statistics and informatics.

Against this backdrop, the future viability of empirical sociology will depend on its ability to adapt to the conditions associated with the ongoing digitization of society. While the new digital technologies provide empirical social research with unique opportunities for data generation and analytical processing, they also impose new methodological challenges that shape research designs, theoretical foundations as well as the methods used. For example, the processes of generating digital process data require the development of tailored measurement and data theories, quality criteria and the corresponding quality assurance procedures to establish shared quality standards besides the still dominating survey perspective. This shift in perspectives is also afflicting how these data are analyzed and interpreted; either by means of traditional statistical models or by means of algorithmic models from the field of machine learning that is receiving increasing attention in sociology.

For this reason, the research topic focuses on (i) big data and (ii) machine learning, which represent two core elements of the developing computational social sciences. Of particular interest are general methodological aspects, the development of corresponding data collection and analytical procedures, quality issues, and innovative empirical applications in a sociological context.

We invite, among others, high-quality submissions on the following topics:

· Evaluation of the quality of big data and methodological challenges in its use

· Methods for dealing with missing data and for ensuring data quality

· The role of research designs in the big data era

· New developments in applying wearables for the continuous collection of process data

· Analysis of high dimensional social data and intensive longitudinal big data

· Evaluation of the validity of inferences drawn from machine learning applications

· Modern analytical strategies for linking big data and classical survey data

· Big data, machine learning and causal inference

· Studies on the validity of machine classifications

· Development of prediction models for social phenomena

· Organizing big social data infrastructures

· Integration of the topics “big data” and “machine learning” into methodological training