BUSY - Building Systems | Education Intelligence

Demo | Get Involved | Architecture | Further Readings

Educational Data Mining (EDM) is a research field that applies data mining, machine learning, and statistical techniques to educational data in order to better understand and improve learning processes and environments. In the context of a software engineering class, EDM can be utilized to analyze data collected from various sources such as student assignments, quizzes or exams. Mining Software Repositories (MSR) is a subfield of software engineering that focuses on analyzing the rich data stored in software repositories, such as version control systems. In the context of a software engineering class, MSR techniques can be applied to study how students engage with the software development process, understand their programming behaviors, identify common issues faced during coding tasks, and assess the effectiveness of teaching methodologies and assignments.

Education Intelligence (EI) is a concept derived from Business Intelligence (BI) principles and methodologies, tailored specifically to the (software engineering) education area. Just as BI aims to gather, analyze, and interpret business data to inform decision-making, with EI we focus on leveraging software engineering data within educational context to enhance and implement a feedback cycle for students and lecturers.

Key aspects of Project “Education Intelligence” include:

Data Gathering and Integration: Similar to BI, Education Intelligence starts with the collection and integration of data from various sources within an educational software engineering class.
Data Analysis and Reporting: EI involves analyzing information to extract meaningful insights. This includes trends in student performance, Git related data, time tracking information, patterns of achieved points, dropout rates, etc.
Ethical Considerations: Just because we can analyze the data from students does not mean we should share our insights with everyone. Results are sensitive student information and can be misinterpreted easily too. At the moment, every information derived based on this project must be treated with caution.
Improving Curriculum Design: Analyze data on student performance and engagement to iteratively improve the curriculum design.

Screenshots & Images

We can provide a short visual overview of our prototype, showcasing its features and functionality. Built on Apache Superset, our prototype embodies Education Intelligence, offering intuitive data visualization and analytics capabilities.

Dashboard

Repository Statistics

One focus of the project is on repository-related data. The simplest form of this is an analysis of Commits over time.

Get Involved

Requirements

Completed course Software Engineering Projekt (PR)
Reliable knowledge in Version Control Systems
Reliable knowledge in Issue Tracking Systems
Optional: Completed course in Software Quality Assurance (VU)

Contact and topic areas

It is possible to accomplish

a bachelor thesis,
a master thesis,
internships,
an ASE PR project

in the context of this project. If you are interested in this project and the following topics please contact: ta.ca.neinullwut.osninull@nullrenbarg.nnnullahoj

Contributors

Johann Grabner, Project lead and Supervisor
Manuel Stöger, Master’s thesis, Initial Development

Architecture

Similar as Project “Binocular”, the architecture is based on MSR principles. Since this project is not aimed at a single repository, but at the entirety of all projects in a semester, an ETL process (Extract-Transform-Load) was implemented with Python. The system is designed to collect, process, and analyze data from software repositories to extract meaningful insights about software development practices.

Components of the Architecture:

Data Sources: GitLab, CSV, git
Data Extraction: A Python script is utilized to extract data from the different sources. Libraries such as GitPython and Pandas are used.
Data Transformation: The extracted data is then transformed in a relational Model to be stored.
Data Loading: The transformed data is then loaded into a central Postgres Database
Data Analysis, Visualization: A modified Apache Superset instance is connected to the central Postgres Database to generate analytical insights through interactive dashboards and charts.