I have been experimenting with ways to add a quantitative dimension to my thesis question. I decided to build a wearable that could measure a person's Galvanic Skin Response (GSR) and send the data to my phone, as well as upload it to a cloud. GSR can approximate a person's arousal state. The idea is that I would walk around different parts of an environment with a user, while the user has my wearable on, and I could correlate physical features with physical arousal state. This would tell me what physical features evoke positive and negative feelings.
Although the wearable "works," it doesn't do the job I was hoping it would do. GSR data is not a very clean dataset to begin with, and it also doesn't quite tell a person's state quite well, as sweat can influence its values a lot.