School of Data Analysis

The Fundamentals of Statistics module aims to provide participants with a comprehensive learning experience extending from data literacy to applied statistical analysis.

Within the programme, participants will learn the role of data in scientific research and core statistical concepts, and will address—through practice—the processes of collecting, organising, transforming, analysing, and visualising data.

During the training, participants will work with data analysis tools such as Excel, SPSS, and R, and examine techniques ranging from descriptive statistics to hypothesis testing, from correlation and regression analysis to multivariate statistical methods.

The module will also address how AI-supported tools can be used for data preparation, facilitating analysis workflows, generating code, and visualisation. Emphasis will be placed on verifying the results produced in these applications, as well as on data security and the reliability of research processes.

Combining theoretical knowledge with applied work, the module aims not only for participants to carry out statistical analyses, but also to select methods appropriate to their research questions, evaluate analysis results critically, and interpret their findings in a meaningful way.

The core purpose of the module is to develop participants’ competencies in working independently with data, conducting statistical analysis, and using their results within scientific research processes.

The Panel Data module aims to equip participants with the skills to analyse social and economic data using the R programming language, apply econometric methods, and interpret findings within a scientific framework. The module offers a comprehensive training process extending from basic programming knowledge to applied econometric analyses conducted with real datasets.

In the first stage of the module, the R programming language and the RStudio environment will be introduced. Participants will learn fundamental R programming structures, data types, functions, and techniques for importing, cleaning, transforming, managing, and visualising data. In this way, the technical foundation required for participants to work independently with large-scale datasets will be established.

Next, the foundations of econometrics and regression analysis will be covered. Cross-sectional, time-series, and panel data structures will be introduced, and topics including simple and multiple linear regression, categorical variables, model specification, interpretation of coefficients, and statistical inference will be addressed. The core approaches of panel data analysis will then be examined, explaining how individual and temporal differences can be accounted for in econometric models.

Subsequently, applications will be carried out with real national and international research data. Comprehensive data sources such as the Türkiye General Social Survey (TGSS), the European Social Survey (ESS), the National Longitudinal Survey of Youth (NLSY), the Health and Retirement Study (HRS), and the Panel Study of Income Dynamics (PSID) will be used. In these applications, participants will experience the processes of formulating research questions, defining variables, preparing data, producing descriptive statistics, estimating econometric models, and interpreting results.

By the end of the module, participants are expected to be able to conduct independent analyses with cross-sectional and panel data, present findings through tables and graphs, and develop reproducible analyses suitable for use in scientific research.

The relationship between the social sciences and data has undergone a profound transformation over the past two decades. A significant part of social life now unfolds in digital environments and leaves traces behind. News archives, social media exchanges, relations among organisations and political actors, and place-based records enable us to examine social processes at a scale and level of detail previously inaccessible. For example, we can track how a wave of protest spreads from city to city through social media posts, and make visible—across millions of texts and interactions—the concepts through which political camps speak and the extent to which they are insulated from one another.

Yet the abundance of data does not by itself produce knowledge. Asking whom these data represent and which voices they leave out, and relating findings to the theoretical accumulation of the social sciences, is more important than ever. Computational Social Science stands precisely at this meeting point as an interdisciplinary field that seeks to forge the link between methodological innovation and theoretical inquiry.

The Computational Social Science module aims to introduce participants who acquired foundations in statistics and Python in the first term of the School of Data Analysis to the principal methods of this field. Module sessions held every Friday from 10:00 to 13:00 over thirteen weeks between 5 March and 4 June 2027 will open with an introductory week discussing the field’s core questions and research designs. In the following three weeks, we will focus on the opportunities and limits that various artificial intelligence algorithms—from machine learning to large language models—offer for social science research. The remaining nine weeks will be organised in three-week blocks devoted to analysing texts with natural language processing techniques, examining social relations through network analysis, and conducting spatial analysis of social phenomena with geographic information systems. In each block, together with specialist academics conducting research in the relevant area, we will combine conceptual discussion with practice on real datasets.

Digital Humanities is an interdisciplinary field that brings together the research questions of the humanities with digital methods, data analysis, and computational approaches. It offers methods for structuring, analysing, and interpreting in digital environments the texts, images, maps, archival documents, manuscripts, and other cultural data produced in fields such as history, literature, linguistics, art history, philosophy, theology, and cultural studies.

This module aims to introduce participants, within an applied framework, to how data are created, organised, analysed, and visualised in humanities research. Alongside the fundamental concepts of digital research processes, the module will provide an applied introduction to methods and approaches such as database creation, text mining, natural language processing, geographic information systems (GIS), network analysis, digital mapping, and data visualisation.

By the end of the module, participants are expected to become familiar with the core methods and approaches of the digital humanities, reassess their own research questions from a data-oriented perspective, and incorporate appropriate digital methods into their research processes in a deliberate and critical manner.

The Psychometrics module is a practice-oriented training programme for researchers in education, psychology, health, and the social sciences who develop or adapt measurement instruments, or who wish to draw on advanced data analysis methods. The module offers an integrated learning process extending from the fundamental stages of scale development and adaptation to structural equation modelling, Rasch models, Item Response Theory, and longitudinal modelling.

In the first part of the programme, scale development and adaptation, expert review and content validity, item analysis, reliability, Exploratory Factor Analysis (EFA), and Confirmatory Factor Analysis (CFA) will be covered. Participants will conduct analyses on real or sample datasets via Jamovi and R Studio, gaining experience in interpreting results and reporting them scientifically.

Structural Equation Modelling (SEM), which attracted strong interest from participants in previous terms, will be expanded this year to cover measurement and structural models, direct and indirect effects, mediation relationships, and measurement invariance applications. Longitudinal Structural Equation Modelling (Latent Growth Models) will also be included, as in the previous year, examining how individuals’ initial levels and patterns of change over time can be modelled.

In the modern psychometrics section, the Rating Scale and Partial Credit models from the Rasch family, as well as two- and three-parameter Item Response Theory models and the Many-Facet Rasch Model, will be addressed. Applications will be carried out using tools such as Winsteps, FACETS, R/RStudio, and web-based applications including Shiny.

The module’s core approach is not merely to show participants “which button to press,” but to enable them to experience which method to use in which situation, what the output means, and how findings can be reported in a defensible manner.

Within the Digital University Platform 2026–2027, a Biostatistics module is offered for researchers in the health sciences, undergraduate and graduate students, and health professionals.

This 10-week programme aims to reinforce the foundations in basic statistics and coding (R / Python) acquired in the first-stage modules through a clinical and epidemiological lens, going beyond the theoretical approaches found in classical textbooks. By focusing on complex study designs encountered in health research, data scenarios that reflect field dynamics, and methodological errors frequently repeated in the literature, it seeks to equip participants with core skills in biostatistics.

What Does the Module Cover?

Application with real research scenarios: End-to-end analysis experience with datasets designed to resemble those you may encounter in a typical clinical or field study.

“Which Test Should I Choose, and When?”: Step-by-step decision-making toward the most appropriate method according to data type, distributional characteristics, and the research question.

Confounding factors and multivariable modelling: Managing confounders specific to health data, interactions, and the correct construction of regression models.

Critical methodological pitfalls: Misconceptions about p-values, overfitting in models, assumption violations, and approaches to missing data.

Everyone who wishes to guide their research with a sound perspective on data and to gain competence in analysis is welcome.

This comprehensive training programme, designed for those who wish to specialise in data-driven decision-making and develop end-to-end analytical projects with the R programming language, consists of four main sections. The first section introduces basic statistics, including data manipulation with R, fundamental probability, and hypothesis testing. In the second section, data visualisation competence is strengthened, and applications in exploratory data analysis, missing data analysis, and outlier detection are demonstrated. The programme then moves into machine learning processes, covering core topics such as regression models, classification algorithms, unsupervised learning, data preprocessing, preventing overfitting, and hyperparameter optimisation. In the final section, more advanced machine learning techniques—including tree-based approaches and ensemble algorithms, anomaly detection, and explainable artificial intelligence (XAI)—are addressed, culminating in the construction of a complete end-to-end analytical data pipeline. The programme is an ideal guide for those who wish to generate value from data and integrate AI models into business processes.