Metis provides a project based approach to learning, teaching end-to-end design, implementation and communication on real-world, complex problems. Topics covered include programming, statistics, data acquisition, machine learning, data visualization, relational and non-relational databases, natural language processing and iterative design.
Incorporated data software tools, such as MongoDB, SQL and Hadoop in tandem with web frameworks like Flask to scale projects on a large level.
Job Recommendation tool
- Collected 10,000 different data science job descriptions from various websites.
- Built an LSI model using NLP libraries -gensim and nltk- to compare keywords, phrasing and n-grams across descriptions in order to produce recommendations.
- Iterated through a process of data processing and feature engineering.
- Implemented Random Forest and Gradient Boosting classification techniques to formulate predictions about where new Airbnb users will travel first.
- Incorporated Pandas and Beautiful Soup python packages to retrieve information for more than 3200 movies.
- Built and optimized linear regression models to predict overall movies success based on a relationship between actor/director characteristics and the overall success of their most recent movie.