Experience Data Scientist brought up on Python, SQL and Tableau. An agile learner and professional who combines exceptional interpersonal skills with strong analytical expertise. Skilled in data acquisition and data modeling, statistical analysis, machine learning, deep learning, recommendation systems, and NLP. With a background in recruiting, real estate, and finance. Strong capacities in mathematics, economics, team building, project management and sales to drive increased revenue and efficiency. Effective collaborator and relationship builder, recognized for versatility, creativity, perspective, sense of humor, and strong work ethic.
- ๐ญ Iโm currently working in the Advertising industry, specifically within Audience creation and insights
- ๐ฑ Iโm currently learning about how to create advanced tableau dashboards to help internal and external stakeholders easily access information
- ๐ฌ Interests include sports, film, media, data analytics (obviously) and reading
- ๐ซ How to reach me: [email protected]
- ๐ Pronouns: He/Him
- Created a machine learning algorithm to predict the outcome of any given NBA game.
- Obtained and analyzed over 23,000 games from the 2003 season and beyond
- Data cleaned and featured engineered through domain knowledge and additional exploratory analysis using Scikit-learn, Seaborn, Matplotlib, Pandas, and NumPy
- Created several different models using various techniques and ultimately obtained an accuracy score of 67%
- Used MovieLens data to recommend movies to any given user
- Retrieved over 100,000 movie ratings from MovieLens and performed exploratory data analysis using Statsmodel, Scikit-learn, Seaborn and Matplotlib
- Created a final model using KNN means and KNN baseline in order to retrieve an RMSE score of 0.8 (ratings on a 5 pt. scale)
- Created a model using Tanzanian water well data from Kaggle to predict which water wells were functional and which were not
- Sourced 59,000 observations on Tanzanian water well data and used exploratory data analysis to better understand features
- Data cleaned initial observations and featured engineered over 100+ new features
- Created 6 different binary classification models, with our final model at an 82% accuracy score.
- Utilized machine learning algorithms to forecast housing prices based on available and engineered features.
- Aggregated data from zillow.com and generated data from niche.com with web scraping.
- Conducted EDA using Seaborn, Numpy, and Scikit-learn for data cleaning and feature engineering
- Utilized linear regression to create a prediction model based on feature selection using wrapper tests.