Konstantinos Kastritis
Computational Physicist | Data Scientist
Bio
Konstantinos is a physicist turned data scientist with a passion for learning and problem solving. He likes exploring any data he can get his hands on and using code to gain insight into real-life problems.
Featured Projects
Sentiment Analysis
Scrape data from reddit using the reddit api and analyze the comment sentiment. Obtain information such as the sentiment breakdown, average sentiment per post, and more.
Stock predictions
Generate predictions for the closing value of securities by using Long Short Term Memory (LSTM) neural networks to model the time series of the stock price.
Brownian dynamics
Brownian dynamics written for gpus, making it easy to scale to tens of thousands of particles for faster results and better statistics on whatever physical system you wish to investigate.
Command Line Music Recommender
Music recommendation program to find new songs. Interfaces with the Spotify dev API, and gives recommendations based on a typical cosine
distance algorithm.
