House Price Prediction
This project predicted the sales price of each house in a neighborhood in Ames, Iowa. I used XGBoost as my model because it has a highly efficient implementation, and I found it has a better performance than others. The most challenging piece of this project would be the data preparation because I felt like I had to keep going back and making changes repeatedly. I was able to place in the top 15% of submissions for this competition on Kaggle, however I would like to keep practicing more techniques for a better score!
NLP Disaster Tweets
The goal of this project was to use natural language processing (NLP) to monitor Twitter Tweets for signals of a natural disaster. I used a process called vectorization to change the words within the tweets into a numerical format that a machine learning algorithm can understand. For the model, I selected Multinomial Naive Bayes to predict wether the tweet contained disaster information or not with 96% accuracy! The most difficult aspect of this project was preprocessing all of the text data to prep it for the model.