This is a repo designated for all projects in this class and other projects I've completed in the past at my time at Bellevue University studying a Masters' in Data Science. Please feel free to look through the content in the folders and see what models and presentations were done in my time here.
Current Projects in the Repo
This project highlights logistic regression in determing wheter a patient's blood markers, physical status, and diet can predict having heart disease.
Given Whole CBC Data (Complete Blood Cell) tests, I utilize Random Cut Machine learning analysis techniques to determine if a patient is at risk of having Diabetes.
Given a dataset from students, we determine positive/negative trends amongst answers provided for satisfaction surveys to further enhance school board's ability to improve metrics for following years curriculums.
With a selection of 2500 cat and dog images, we are able to train a model using an Image Data generator to classify these images to flattened binary files. They are then tested on single images to determine if they are indeed a cat or dog image.
Given a dataset of steam sales from game releases, we attempt to predict prices for games by only knowing their initial features of the game.
Given images of pizza and images that are not pizza - We predict if given images are pizza or not!
In this dataset, we are given tweets and the provided gender of the twitter user. We attempt at determining if we can predict the user's gender off these tweets and find useful features that may seperate the two if there are any.
This script predicts IMDb ratings for episodes of "The Office" using a Linear Regression model. It preprocesses episode descriptions with TF-IDF vectorization and combines them with numerical features like season and episode number to train and evaluate the model.
We make a bar chart shows how many teams ended up in each postseason ranking category, giving us a sense of the distribution. A correlation heatmap reveals which features are most related to postseason success. A scatter plot compares predicted rankings with actual ones, visually checking our model's accuracy; and a bar chart highlights the importance of each feature in predicting rankings.
I take a large excel of video games done by Nintendo Switch and provide cool visualization with explanations!