This is my typically late review of the past semester, it’s hit home now that this is the penultimate one of these that I’ll be writing, and the end of University is starting to feel real now - but the future is exciting.
The past semester has been incredibly busy and hasn’t stopped, but I’ve been up to some cool stuff along the way.
Data Mining was my favourite module of the semester and looked at what data mining is, how it works, and why it’s incredibly useful. As well as learning the fundamentals of how it works mathematically and how the tech has progressed to use deep neural networks from the simple neurons used historically, we applied the techniques to a real-world context in the coursework. Through Jupyter notebooks and using pandas to clean the data and spot patterns, building different types of models using scikit-learn.
We applied this to a medical context in the coursework using anonymised health data showing various cardiac events and whether they lead to a heart attack. The coursework had an interesting problem domain, as the justification of prediction methods wasn’t just focussed on what provided the highest accuracy metrics, you had to consider the potential of misdiagnoses and how this would affect people.
Overall, it was a lot of fun, and I’d recommend having an experiment with data mining notebooks. After exams, I’m planning to explore my student data record from the myEngagement project and see what new information I can find and visualise after cleaning up the dataset.
Virtual Environments was an interesting module as it looked at how 3D works behind the scenes. As someone not studying games modules, it gives an overview of how artificial worlds are created and made to be believable. It was also fascinating from a psychological perspective to understand how the brain fills-in gaps from what it sees that allow the creation of these experiences.
For the coursework in the module we were tasked with conducting a scientific experiment to investigate how either VR and AR affect peoples understanding of their surrounding, in my group we used Microsoft HoloLens gaming, combined with CS:GO and used stroop tests and found a correlation in reduced reaction time with users physically immersed in a game as opposed to playing it on a PC.
The dissertation is coming along well, and I’ve just had my mid-project progress review to assess how far along with it I am. Currently, I’ve built my API with the serverless framework, and have it deployed in production on AWS, using CircleCI to run my build scripts.
SwiftUI turned out to be a cool challenge, with some strange and undocumented bugs along the way, but I’m fighting through them! Overall, as a way to develop apps, I’m liking SwiftUIs style. Just got the rest of this next semester to complete everything and write it up now though.
I’m looking forward to sharing more about how my project has developed, and what I’ve learned along the way after submission is all complete in April.