Welcome to the coding page. This is where I'll post updates about my coding projects. I'm not sure how elaborate or detailed I want this to be, it's really a sort of habit tracker more than anything else because I want to start practicing coding every day. The 100 Day Challenge is quite popular on github and tech groups. It's as the name suggests: coding a project everyday for 100 days. I think this'll be good for me because when I sit down to do any project I always have the expectation of it being this Great Big Thing, when in the first place little projects need to exists to create the foundations of big projects. Or whatever. Anyways I'll be logging these projects here! I don't think I'll be showing any real code, atleast not for now, but I'm going to give detailed descriptions to try test and solidify my own understanding of what I'm doing. Hmmm...
February 17 2025
100 DAY CHALLENGE! DAY 2: Today I was helping a friend out with some NLP. Part of their project is training
a model to correctyl detect emotion from text. I did something similar with reccomending movies during my exchange, so it's not
new to me, but I wanted to try my hand at this task too. The basic model began with a 33% accuracy, which obviously isn't amazing
but I fiddled around with the values and the algorithms for a few hours and I was able to get the accuracy up to
68%. An ideal accuracy for training models is a minimum of 70%, so I need to work on it some more. My model can accurately
destect the emotions from many sentences, but one issue I came across is it sometimes attributes emotions to words that are completely neutral
without context. For example, sentences with the word "college" keep returning "Sadness" even if they are positive, which is a bit funny.
It's probably my data. The dataset is just something I pulled quickly from a dataset site, so it is probably not
long enough or detailed enough for my model. Will continue messing around with it.
but
February 16 2025
100 DAY CHALLENGE! DAY 1: Today I made a simple web scraper for wikipedia articles.
The program asks the user for a topic and spits out the introduction of the respective
Wikipedia page. This took me about 15 minutes. I want to come back to it though. It comes
back empty handed for a lot of articles, probably because of the difference in formats,
so I need to fix that. I also want to try put it to use maybe for some sort of quiz game or a
mini search tool. I don't know. We'll see.