Pneumonia classification from chest x-ray images (Feb 2021 - Present)
Working with the Chest X-Ray Images (Pneumonia) dataset
The dataset consists of 5,865 chest X-rays of various resolution and aspect ratio. They are classified into sets that are normal, have bacterial pneumonia, or have viral pneumonia. Here are some example chest X-rays.
Project is in its early stages. See the current writeup and jupyter notebook for more details.
Spotify song populartity regression (Feb 2021)
Working with the Kaggle Spotify Dataset 1921-2020, 160k+ Tracks dataset
Trained RandomForest, DNN, LinearRegression, and LinearSVR to predict songs populartiy. Obtained a RMSE error of 12.300 which is nearly 2x more accurate than guessing the average popularity. See the writeup and jupyter notebook for more details.
