I created a Classification Model that predicts probability whether a country will default on its debts or not in the following years.Analysis of how features such as inflation, exports, and Government Expenditures % of GDP affectlikelihood of Sovereign Default and Restructuring.I developed a Web App allowing users to explore model, adjust features. Intended to help investorsavoid buying bonds in countries likely to default, and understand what factors lead to Sovereign Default.Data was economic data from the IMF since 1980 and Historical Sovereign Default and Restructuringdataset since 1970.Tools: Binary Classification, Supervised Learning, Web App. Matplotlib, Numpy, Pandas, Scikit-learn,Seaborn. Tableau for visualizations. XGBoost, Random Forest, KNN,Naive Bayes, Logistic Regression. Streamlit for Web App. Hosted on HerokuBest results were achieved with Logistic Regression. Debt/GDP seems to be most correlated with acountry defaulting on its bonds.

Analyzed how topics in the State of the Union addresses have changed over more than 200 years of American history and compared which speeches were most similar topically using Natural Language Processing (1790-2018)I used data pulled from Kaggle on the State of the Union corpusTools: Python Pickle, NLTK, Pandas, Scikit-learn, Matplotlib. Count Vectorizer, TF-IDF, NMFTableau used for plots in presentation slides

Created a model for a credit card company to avoid issuing credit cards to users who default on their debts.Pulled data from a Kaggle data set consisting of anonymized credit card data of clientsTools: Matplotlib, Numpy, Pandas, Scikit-learn, Seaborn, XGBoost, Random Forest, KNN, Naive Bayes,Logistic RegressionAt this time, I also did work on SQL queries. This improved my personal confidence onStructured Query Language

Utilized regression to find which features are correlated with the highest Movie Box Office returnsData was pulled from Box Office stats on Top Lifetime Grossing movies and was analyzed after webscraping text from BoxOfficeMojo.comCreated features from scraped HTML including: Distributor (studio), Budget (USD), Foreign MarketCount, Genres (10), and Rating (MPAA)Tools: Pandas, Numpy, Scikit-learn, Seaborn, Matplotlib, BeautifulSoup, LXML

Provided recommendations on optimizing deployment of a tech client's team to collect email addressesof potential attendees for their upcoming gala eventTurnstile data from New York City subway stations (MTA) in combination with NYC demographic datafrom the census was used for the projectTools: Pandas, Numpy, Scikit-learn, Matplotlib, and SeabornIdentified five MTA stations with the most traffic over our 4 week periodThe data was aggregated to find the average weekly activity based on time and weekday anddemographics were analyzed for the top five stations.

Daycare for children ages four to ten.I supervised camp activities to ensure children's safety.Experience communicating any updates or issues with supervisors on staff.Collaborated with other counselors to maintain a fun and safe environment for the children.

Daycare for children ages four to ten.I supervised camp activities to ensure children's safety.Experience communicating any updates or issues with supervisors on staff.Collaborated with other counselors to maintain a fun and safe environment for the children.