Anomalies Detection in Student Activity in Solving Unique Programming Exercises: Motivated Students against Suspicious Ones

The proposed dataset contains messages from the Digital Teaching Assistant (DTA) system, which captures the results of automatic verification of students’ solutions of unique programming exercises for 11 tasks of various types, which are automatically generated by the same system that automates a massive Python programming course at MIREA - Russian Technological University (RTU MIREA). The dataset contains anonymous information about students, their groups, variants, the time of successful and unsuccessful attempts to submit solutions to exercises for 11 tasks to the DTA system, as well as the ways in which a particular student performed a particular unique exercise. The dataset can be the subject of exploratory analysis in terms of detecting various anomalies and outbursts both in the structure of a multidimensional time series formed from messages sent to the DTA system as a whole, and in the behavior models of student groups and individual students when they interact with the DTA system.