Projects
Cycles of Demand: Predicting Washington DC's E-Bike Rentals Usage
Regression | EDA
Understanding user patterns and trends is vital for improving business strategy and customer satisfaction.
In this project we explore user behavior around E-bike usage and use machine learning to forecast the number of riders per hour based on date, time, and weather conditions. We seek to understand why users ride when they do, and propose business changes to increase operational efficiency and customer satisfaction.
random forest regression, linear regression, lasso regression, variable selection, one-hot encoding, standard scalar, cross-validation, R^2, RMSE, sklearn, seaborn, matplotlib
Technologies:
Automated Grading of Baseball Trading Cards using Convolutional Neural Networks
Computer Vision | EDA
Can you predict how good wine will taste based off on specific physical and chemical characteristics? What makes one wine better than another?
In this project we try to answer these questions and more.
python, pandas, scipy, matplotlib, sklearn, seaborn, xgboost, random forest, naive bayes, logistic regression, randomizedsearch, gridsearch, roc, auc, f1 score, feature importance
Technologies:
Predicting Wine Quality From Physical and Chemical Properties
Classification | EDA
Can you predict how good wine will taste based off on specific physical and chemical characteristics? What makes one wine better than another?
In this project we try to answer these questions and more.
xgboost, random forest, naive bayes, logistic regression, randomizedsearch, gridsearch, roc, auc, f1 score, feature importance, scipy, matplotlib, sklearn, seaborn
Technologies:
Your Personal Data Science Career Coach
LLM Chatbot
Do you have a favorite podcast? Wouldn't it be nice to have a chatbot that already had listened to all of the episodes?
In this project I build a chatbot that knows everything about The Analytics Power Hour podcast. Hosted by top data scientists and analytics directors, this podcast covers data science career progression, best practices, and distils nuggets of data science wisdom from the industry leaders that they interview.
GPT3.5-turbo api, steamlit, prompt engineering, custom embeddings, retrieval augmented generation (rag), large text summarization