Here you can find some of the projects I've worked on.
I developed a machine learning model to detect stress from social media posts. The project used datasets from Reddit and Twitter to train and evaluate the model. I compared the performance of transformer-based neural network (BERT) and traditional machine learning (logistic regression) for stress detection from textual data.
I developed a machine learning model, which was trained on a dataset of smartphone sensor data. Two machine learning methods were used: logistic regression and decision trees. I then compared the performance of the two methods. Both of the methods were able to predict user activity from unseen smartphone sensory data with a good accuracy.
See more: Report
As a part of my bachelor's thesis, I developed a framework to evaluate decision-making in economic games. The framework differed from pre-existing frameworks in that it considered interactions with non-human agents, such as AI players. My thesis as a whole considered the use of game-theory as a tool to evaluate economic decision-making in Human-Computer Interaction.
See more: My bachelor's thesis
We conducted a questionnaire study with my fantastic group to investigate the relationship of psychological effects and social media usage on students. We used a survey to collect the data and analyzed it using R. The study was presented to professors and doctoral researchers in Aalto University. The report is currently only available in Finnish.
See more: Report
As the widely used Discord bot "Rhythm" that usually plays music on voice channels was down, I did a small project to develop a Discord bot that plays music by downloading the audio from Youtube and playing it online similarly as the "Rhythm" bot. You could prompt the bot a natural language search query and it automatically found the most relevant audio from Youtube and played it.
See more: Github
I created a data-visualization library for numerical json-formatted data. The library formed a graph that represented the inputted data and automatically adjusted the graph's range and scale for values depending on the data. Users could also name the graph axis and add a title to the graph.
