Below you will find a list of topics and ideas for student projects - particularly bachelor and master theses - for students Computer Science and Information Sciences at the Radboud Universiteit. Please also have a look at the Master Thesis Lab for formalities and general instruction. There is also an overview of external and internal thesis topics and opportunities for Information Sciences.
In addition to the topics and ideas below, I am interested in various other topics related to my research focus. Have a look at my publications for inspiration or simply make an appointment.
Some general topics that I am currently interested in
User Modeling and Personalization:
- Privacy-Aware Recommender Systems
- Mining Social Media Profiles
- Detecting Bias in the Media
- Echo Chambers and the Filter Bubble
Build your own filter bubble!
What will happen if users will build filter bubbles themselves and what can we learn from this in terms of user preferences, satisfaction and overall results? Which dimensions do users apply and in which order? How do they respond to the resulting filtered content - as expected? Do they need more information? At the end, are they able to explain and are they happy with the results?
Why was this ad shown to me?
Personalized ads are often considered creepy, either because users directly can relate them to previous activities ('why does the system know that?') or because they seem the result of the user being put in a particular user profile group (based on interests, demographics, …).
In this project, you are expected to conduct a user study in which you collect:
- Ratings on 'creepiness' for different types of actual ads and explanations why users consider this ad creepy
- Ratings on how satisfying the users find the explanations provided to them and explanations behind the ratings
The goal of this project is to design guidelines for user-friendly, transparent explanation mechanisms for advertisements that would reduce creepiness.
The quantified self: do users want data to control their lives?
An increasing number of people actively count or measure their steps, exercises, food intake and other measurable aspects of their daily lives, health and well-being. Products like Jawbone and Fitbit promise that they will provide incentives for a better, happier life, but:
- To what extent do users want data and recommendations to control their lives (or to what extent do they think they need it)?
- How sensitive do they consider the data (are they happy to share it with doctors, close family, or advertising companies)?
Mutual privacy and good manners on the social web
There are topics that you would discuss with your partner, but probably not with your colleagues. If you are in a pub and you want to discuss a more confidential or private remark, you typically lower your voice or otherwise ensure that not everyone can hear it. We do not only behave like this to protect our privacy, but also to avoid that other people might feel embarrased by learning facts that they might consider embarrassing.
In platforms such as Facebook, there are privacy mechanisms to determine your audience (e.g. friends, friends-of-friends, public). On Twitter, public tweets are the norm. WhatsApp is for direct conversations with people that you know in person, individual or in groups. In Snapchat, messages disappear after a while. This leads to questions like:
- Which platforms do we choose for which type of communication? What drives these choices (convenience, availability, security, privacy, ...)?
- Do we behave differently in different platforms? Are there differences in social norms? What types of situations do users consider embarrasing?
- Are there any differences that can be related to demographics (like age group, gender, education level, nationality, ...)
