Portfolio of data science projects completed by me for learning purposes.
Computer Vision
- One Man's Trash: One-Mans-Trash is a web application that uses deep learning techniques to detect and classify different types of trash objects in images, such as plastic bottles, cans, cups, and plastic bags.
Natural Language Processing
- Finance Complaint Classifier: The Finance Complaint Classifier is a deep learning model designed to identify the financial produc (credit card, loan, mortgage, debt, etc.) that a customer is complaining about based on their complaint narrative.
- SkimLit: Deep learning NLP model used for sequential sentence classification of medical paper abstracts. Classifies the sentences of a medical paper's abstract in to the following classes: background, objective, method, result, or conclusion.
Machine Learning
- Superstore Marketing Campaign: A series of exploratory data analysis and model experimentation to aid the marketing campaign of a deal that classifies a customer on their likelihood of availing for the deal.
- Beijing Pollution: Time series forecasting on pollution levels of Beijing given different variables like temperature, humidity, etc.
- Concrete Compressive Strength: Regression modelling on the strength of concrete given different variations in their formulas.