DISA

When? Friday February 2th 14-16
Where? Onsite: D2272 at Linnaeus University in Växjö and online
Registration: Please sign up for the PhD-seminar via this link https://forms.gle/H9oqD7duf23ibLfv7 by January 31st (especially important if you plan on attending onsite so we have fika for everyone)

14.00-14.10 Welcome and practical information from Welf Löwe
14.10-14.55 Presentation and discussion: Object Identification in Land Parcels using a Machine Learning Approach – Niels Gundermann, Industry PhD student at Data Experts
14.55 – 15.05 Coffee break
15.05 – 15.50 Presentation and discussion – Development of Visual Learning Analytic Tools to Explore Performance and Engagement of Students in Both Primary/Secondary and Higher Education – Zeynab Mohseni, PhD-student
15.50 -16.00 Sum up and plan for our next seminar on February 2nd

Abstracts

Object Identification in Land Parcels using a Machine Learning Approach – Niels Gundermann, Industry PhD student at Data Experts

Since change detection on aerial images is an important subject for cadastral work, we introduce an AI based approach to the detection of human made objects on land parcels. To this end, we use binary image classification performed by a convolutional neural network. Binary classification requires the selection of a decision boundary and we provide a deterministic method for this selection. Also,we varied different parameters to improve the performance of our approach leading to a true positive rate of 91.3% and a true negative rate of 63.0%. A specific application of our work supports the administration of agricultural land parcels eligible for subsidies. As an effect of our results, the authorities could reduce the effort for the detection of human made changes by approximately 50%.

Development of Visual Learning Analytic Tools to Explore Performance and Engagement of Students in Both Primary/Secondary and Higher Education – Zeynab Mohseni, PhD-student

There is a significant amount of data available about students and their learning including text, grades, quizzes, timestamps, and behavioural data in many educational systems today. However, this data is often spread across different digital tools, making its strategic use challenging. Moreover, there are no established standards for collecting, processing, analyzing, and presenting this data. This lack hinders teachers and students from making informed decisions, presenting a significant obstacle to progress in schools and effective development of Educational Technology (EdTech) products. Visual Learning Analytics (VLA) tools show promise in improving decisions in primary and secondary schools, but their effectiveness has limited evidence. This project aims to co-design and develop a VLA tool for analyzing students’ performance and engagement, employing Machine Learning and Data Mining techniques to create a decision-support tool for teachers.

Collaborating with four Swedish municipalities and four EdTech companies producing Digital Learning Materials (DLMs) allowed leveraging their data to enhance digital teaching methods. However, the absence of a common data standard for data poses challenges for replication across municipalities and EdTech companies. To address this, we proposed a technical solution involving a data pipeline using the secure Swedish warehouse, SUNET, to facilitate information exchange. This includes a data standard for integrating educational data into SUNET, along with customized scripts to reformat, merge, and hash student data. Implementing this data standard and technical solution promises improved management, transportation, analysis, and visualization of educational data for diverse stakeholders. This doctoral project also introduces a scenario-based framework for VLA tool development, highlighting the crucial role of user involvement. Additionally, we proposed a Human-Centered Design approach engaging teachers in co-designing a simple VLA tool. Prototyping, guided by learning and visualization theories, ensures a tool that understands and analyzes educational data, monitors student learning paths, simplifies tasks, and gives teachers more time for teaching and growth.