About Me

Hi everyone, I am Gabby! I am a recent Computer Science graduate from NYU and a Data Science Immersive bootcamp. I am interested in data science and artificial intelligence, specifically data visualization, natural language processing, machine learning, deep learning, and computer vision. As for general interests, I love reading, playing video games, binge watching shows, trying new foods, and traveling. I also love spending time with my friends, family and my dog, Snoopy (picutred here)! I hope you enjoy exploring my protfolio and more of my projects on Github. My contact information is at the bottom! Thank you :).

The One With Natural Langauge Processing

This was a fun project because I love Friends and I like NLP. I used natural language processing and machine learning to determine which character said the lines. I first took a dataset I found on Github with the season and episode the line was said along with who said it(there were more features, but I only used these). Then, I did some data cleaning, only taking the lines from the main characters (Rachel, Ross, Monica, Chandler, Pheobe, and Joey) and deleting one word dialogues to try to better the models. Then I did some exploratory analysis like what were common words, how many lines each character had throughout the whole show and per season, and more! Then I created five different multiclassification models, Random Forest, Logistic Regression, K Nearest Neighbors, and Näive Bayes, and Ada Boost. The Näive Bayes and Logistic Regression models did the best, while the KNN and Ada Boost models did the worst. I also created a binary classification model for only Rachel and Ross, and Monica and Chandler. I created a web app through Streamlit so non-technical users can explore and have some fun!

Brain Tumor Detection

For this particular project, I used Google Colab.

In this project, I used the Kaggle dataset for brain MRI images. There were a yes folder and a no folder. The yes folder contains all the MRIs that have tumors, and the no folder contains all the MRIs with no tumor. For my models, I used convolutional neural networks in order to detect if an image has a tumor or not. I created different functions to read and convert the images, to plot the accuracy of the training and testing sets, building the CNNs, and fitting the CNNs. I decided to do this project to learn and teach myself about image classification and how CNNs work.

Puzzle Game

This program was written in Java and it finds all the solutions to a number puzzle (represented by an array of positive integers). In each path, there will be directions (left or right) to follow the path it found. If there is no solution, then there was no path found. There can be multiple solutions. This project uses recursion in order to find these.