Construction of university student portraits based on smart card data
Zhao, Kai (2025)
Kandidaatintyö
2025
School of Engineering Science, Tietotekniikka
Kaikki oikeudet pidätetään.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi-fe2025052253051
https://urn.fi/URN:NBN:fi-fe2025052253051
Tiivistelmä
The rising digitalization of university campuses has led in the collection of massive amounts of smart card data, which records students' daily activities such as dining, library use, and dormitory access. Effectively analysing this data can support personalized education and student-centred management. This thesis aims to construct student portraits based on smart card data, enabling education administrators to gain insights into behavioural patterns and make data-informed decisions.
In this research, a data-driven student portrait system was developed using Python, Flask, HTML5, JavaScript, and ECharts. The process involved preprocessing raw smart card data, designing a label system for behavioural attributes, and applying K-Means and GMM clustering algorithms to group students with similar habits. A visualization platform was built to present clustering results, consumption patterns, and individualized behavioural labels through interactive charts and word clouds. System testing confirmed the correctness, responsiveness, and scalability of the platform. The results demonstrate the feasibility of smart card-based behavioural analysis and its potential for improving educational services.
In this research, a data-driven student portrait system was developed using Python, Flask, HTML5, JavaScript, and ECharts. The process involved preprocessing raw smart card data, designing a label system for behavioural attributes, and applying K-Means and GMM clustering algorithms to group students with similar habits. A visualization platform was built to present clustering results, consumption patterns, and individualized behavioural labels through interactive charts and word clouds. System testing confirmed the correctness, responsiveness, and scalability of the platform. The results demonstrate the feasibility of smart card-based behavioural analysis and its potential for improving educational services.