Data Science Immersive Fellow
DATA SCIENCE PROJECTS:PREDICTING HEART DISEASE & DIABETES USING BINARY CLASSIFICATIONGathered, cleaned and processed health survey data from the CDC. Created binary classification model that predicts likelihood of heart disease & diabetes with 83% accuracy. Created Streamlit app that takes user input and suggests resources & informed discussion with a physicianPREDICTING LIKELIHOOD OF LOAN SUCCESS USING NLP and GRADIENT BOOST CLASSIFIER Worked on a 4-person team to develop model that improved success likelihood for Kiva loans. Used binary classification and NLP to create machine learning model with 83% accuracy. Created Streamlit app to be used in-house at Kiva to help improve loan pre-screening processUSING NATURAL LANGUAGE PROCESSING TO DETERMINE SUBREDDIT ASSOCIATION Used Pushshift API to gather data from Popheads and Let’s Talk Music subreddits. Cleaned and processed data using Regular Expressions, WordNetLemmatizer & CountVectorizer. Final model: AdaBoost Classifier with 85% accuracy.