Studying Business Information Systems Engineering

Studying Business Information Systems Engineering

Topics for BSc./MSc. Theses

In the course of your studies, you will have to write a master's thesis and/or bachelor's thesis, depending on the course of study. During this time, you will acquire new specialist knowledge and learn the tools of independent development and research work under the guidance and supervision of a supervisor of your choice. Your first step towards successful work is to find an interesting topic for you. To do this, contact a research assistant who could supervise you. In your search for the right supervisor, please orient yourself on the subject area you are looking for and the research focus of the individual team members. Develop as precise an idea as possible of the topic you are interested in. Below you will find the current range of topics offered by the Chair of Information Systems. The list can be extended - in individual discussions with the academic staff.

The list below is not complete! Please also see the german descriptions if possible. Of course you can also always write the research staff of the institute for vacant research topics.

Bachelor Theses

Estimating Human Energy based on EDA Data

Supervisor:
Fabienne Lambusch,
Fraunhofer IGD

Suitable for: bachelor thesis

Language: German/English

Topic Description

Human energy has become an important topic in work and organizational psychology. It is an umbrella term for sensations like feeling vital, awake, and concentrated and characterizes the energy state of a person as a form of personal resource status. Strain and recovery are reflected in a respective decrease or increase of the individual energy level. A high energy level plays an essential role in the employees' ability to act and their motivation to accomplish their work. Furthermore, long-term psychological strain can cause serious clinical pictures, such as burnout that is associated with a low energy level. In order to prevent such negative consequences and promote a sustainable management of personal resources in an organization, it is vital to support employees individually in reflecting on their energy status and in adapting their behavior.

However, the options for obtaining information about the energetic wellbeing of a person are so far limited to highly subjective and time consuming methods such as self-assessment questionnaires. Especially in the context of corporate health management, there is a need for objective and quick methods with available devices to determine the level of human energy. Sensor-equipped wristbands like the Empatica E4 enable monitoring vital signs like heart rate or nervous system activity. Electrodermal activity (EDA), also known as galvanic skin response (GSR), is already associated with human emotional states or stress. Thus, it could also be a reasonable basis for human energy estimation in the workplace. This bachelor thesis aims to analyse human energy and Empatica E4 data from prior experiments (already existing data), especially EDA data, in order to test the possibilities of estimating the human energy level with these. For evaluating whether the data could be used for a reliable, objective measure of the energy level, the main part of the thesis is to test and compare different approaches for finding connections between concepts e.g. from simple correlations to different statistical models.

Performance-Rating of Free-to-Use Triples-Stores

Supervisor: Achim Reiz

Suitable for: Bachelor

Language: English (prefered) or German

Date: 19.02.2020

Semantic services as a symbol-based artificial intelligence technology can be used in various application scenarios, e.g., enhancing deep-learning systems, autonomous driving, or knowledge-based systems. With SPARQL, there is a widely accepted query language standard available, allowing uniform access regardless of the underlying technology - similar to the SQL-standard for relational databases. But while the landscape of relational storage is highly mature and almost all SQL-based storage systems have high performances, this is not the case for triple-based databases. These, often research-driven developments, are more diverse regarding their functionalities and efficiency. This makes the selection of the right tool for the selected goals complex.
Your task is the evaluation of free-to-use RDF-based triple-stores like apache fuseki or virtuoso. Using uniform queries, you will measure the performance of the software and compare the expected results. Based on your findings, you can derive recommendations which database software is best to use for which application case.

Semantic Augmentation for Image Recognition

Supervisor: Achim Reiz

Suitable for: Bachelor or Master

Language: English (prefered) or German

Date: 05.01.2021

 

Open Images is a popular open-source dataset comprising of about 9 million annotated images. Developers can use it to train and set up domain-independent object recognition systems. With TensorFlow Hub, there is already a pre-trained neural network (NN) publicly available that can easily be integrated into existing applications, bypassing the need to train a new NN using the 550 Gb large Open Images repository. While the trained object detection delivers promising results, the detected items are isolated, and a shared context is missing. A high-level semantic is missing, allowing the interpretation of the results beyond the elements that are visible. Such can be achieved using ontology methods.

In 2020 we published SemOI - Semantic Augmentation for Open Images (The prototype is available on http://semoi.informatik.uni-rostock.de/). However, due to rather granular relations between the concepts, the accuracy of the detection is low. Your task is to improve the augmentation by revising the underlying ontology. Your first task is the definition of an augmentation goal and the corresponding high-level concepts. For example, one goal could be the detection of situations: A "car" and "tool" detected together could indicate a "repair"-situation. The next step is the modeling of these relationships. A performance evaluation concludes the research.

Standard Software for Participatory Budgeting (PB)

Supervisor: Achim Reiz

Suitable for: Bachelor

Language: English (prefered) or German

Date: 23.02.2021

 

Participatory Budgeting (PB) allocates a percentage of an administrations' budget for the direct democratic distribution. At the first stage, Citizens can propose new ideas, like park benches or an improvement of a playground. In the next step, all citizens can vote for the proposals they support the most. PB allows to use the limited financial resources of a municipality more efficiently and engages citizens in political decision-making, thus has the potential to strengthen democratic processes.

The use of IT-technologies often drives these initiatives by providing forms for uploading ideas, online forums for discussions, and virtual voting facilities. While municipalities often develop customized solutions, some standard software is available, like DecidimConsul, or citizens.is

Your task is a comparison of standard software for PB. The research shall begin with an extensive search in the internet and literature databases to find relevant software and their license, pricing model, and availability. In the next step, your task is the comparison of the identified software: What kind of features do they implement? What are the differences between these artifacts?

The research concludes with a software recommendation based on your identified advantages and shortcomings. 

Master Theses

Visualizing Ontology Evolution

Supervisor: Achim Reiz

Suitable for: Bachelor or Master

Language: English (prefered) or German

Date: 20.2.2020

Computational Ontologies offer a means for automated knowledge generation and have the potential to be a cornerstone for the upcoming challenges in the field of artificial intelligence.  
Their increased usage leads to a rising need for evaluation.
The OntoMetrics-tool (ontometricts.informatik.uni-rostock.de) of Rostock University provides a web interface for the calculation of ontology metrics. Recently, this tool was extended with a feature to analyze historical data based on git-repositories.
Currently, the tool does not offer advanced visual analysis capabilities. We argue that visualizing metric developments have the potential to significantly improve the intuitive understanding of the evolutional aspects of the ontology.
Your task is to research creative new ways to encode the analyzed data visually. As a preliminary step, the current state of the art regarding ontology visualization has to be collected. In the next step, you will build a front-end that enables the dynamic creation of diagrams based on a visualizing framework of your choosing (e.g., d3.js). These diagrams shall
1. provide an intuitive visualization of the metric development
2. set various variables into context with each other
3. be downloadable as a picture or pdf.
The research concludes with a validation of the usefulness of the newly created approach through a user-study.

Integration of Sensor Data in Electronic Diary Studies for Personal Feedback Generation

Supervisor: Michael Fellmann

Suitable for: Master Thesis

Language: English/German

Earliest start time: February 2022

Topic Description

IT-supported diary studies enable to capture and analyse a plethora of data concerning e.g. work-related activities as well as aspects from other life domains. This can lead to improved self-reflection or even ignite behaviour changes leading to increased performance and wellbeing. Recently, diary studies with personal feedback have been developed in Rostock that provide users with personal feedback immediately after study completion. However, such studies up to now rely on manual data entry which is a tedious task. While for some scales like work progress, manual data entry seems to be inevitable to ensure accurate data, other scales such as heart rate or computer-based work could be recorded in an automated way. Since an integration of such sensor data is lacking in current diary study tools, the goal of the master thesis is to integrate such data into IT-supported diary studies. On the conceptual level, the task is to develop an approach how to aggregate fine-granular time series data and combine it with diary study data. The approach should also encompass a method for data fusion, i.e. when and how to fuse sensor data and diary study data. On the technological level, a proof-of-concept that showcases this integration should be instantiated. In this direction, a time series database is available (InfluxDB) and should be used in conjunction with the formR diary study tool, both hosted at the Institute of Computer Science. For the evaluation, a simple real-life dataset that will be provided by a smartwatch and/or a desktop activity logger should be used along with a simple diary study that showcases the joint analysis of data from different sources inside the generated personal feedback.

Extension of a Desktop Activity Logger with a Recommendation Widget

Supervisor: Michael Fellmann

Suitable for: Master thesis

Language: English/German

Earliest start date: February 2022

Themenstellung

Desktop activity logging aims at recording data about IT-supported work and life activities performed using a (personal) computer. The data is then used for documentation or self-reflection that can lead to insights. Such insights can comprise simple facts such the preferred working hours of the user or which applications are used frequently. More sophisticated analyses could determine e.g. work-related productivity, distraction or even flow state. Despite the plethora of data that is captured, state of the art logging tools do not help the user to accomplish his tasks via recommendations. Such recommendations could comprise e.g. context-relevant files, folders, applications or websites that might be useful in the respective work situation as evidenced by previously recorded data. Beyond that, the recommendations could be complemented by further tips on e.g. taking regular breaks, making gymnastics or mind-set-related hints for a positive and engaged work attitude. The goal of the Master Thesis is to design, implement and evaluate a recommendation widget that generates useful recommendations for the user. The conceptual part will be to develop a model for context-sensitivity and automated adaptation of the recommendations to the user preferences. The technological part will be to extend an existing desktop logger implemented in C#. Finally, technical criteria as well as initial user feedback e.g. regarding the perceived utility of the recommendations should serve for a preliminary evaluation.