Data validity is one of the fundamental concepts determining whether data accurately describe the real world. Despite the abundance of modern databases, official statistics in the post-Soviet space do not always reflect reality, making their use for analysis challenging. This course aims to introduce students to basic methods for verifying the validity of quantitative data, as well as advanced techniques for constructing alternative variables that better describe underlying phenomena.
The course consists of two parts. The first part focuses on the key steps of data validity analysis: understanding the data-generating process, evaluating data quality, identifying and handling outliers, and improving data quality before applying analytical methods. This part of the course covers tools such as Benford’s law, terminal digit analysis, and related methods.
The second part focuses on regional data, which often contain manipulation. It introduces modern methods of geospatial data analysis for the social sciences and aims to develop skills in geospatial visualization, identifying spatially driven relationships in data, and constructing alternative indicators when official data sources may lack objective or reliable information.
The course uses modern free software to ensure reproducible and interpretable analytical results. It also includes group-based geospatial projects designed to foster collaboration among participants.
The course literature includes recent applied research published in leading international peer-reviewed journals. A wide range of real-world data examples will be explored. By the end of the course, students will have acquired essential data quality assessment skills for their future professional or academic work.