A GUI for visualizing player shot data for any season post-2015.
Users can create Heatmaps, Shot Charts, Bar Charts, and more for NBA players.
The comparison tool allows users to quickly inspect similar players found through KMeans clustering
A deep dive into the comments section of a popular YouTube channel. In this project, I used python to gather/explore the data and investigated the relationship between sentiment and other video metrics.
I extracted, cleaned, and analyzed the Zillow listing data for 4 major cities in Minnesota with the goal of answering questions such as:
- Which listing features have the strongest correlations with price?
- How many listings in each city have decreased from the original asking price?
Medical abstracts can often be unstructured and difficult to read. The objective for this project was to create a model for classifying sentences as part of the Objective, Background, Methods, Results, or Conclusions section of the document to improve comprehension.
I fine-tuned a Vision Image Transformer model to take in an image of a bean leaf and predict whether it is healthy or diseased. After training the model and achieving 96% accuracy, I created a Gradio web-application for users to classify their own images!
I used python to explore responses to Kaggle's annual survey of data professionals and predicted salaries with a Random Forest, KNN, SVR, and a VotingRegressor.
Calculating cup-to-disc is a manual task for a professional that could be improved and made more consistent with deep learning. In this project, I used 2 popular pre-trained neural networks (EfficientNet and VGG16) along with regularization through image augmentation and Dropout layers to detect glaucoma in fundus images.