Jeffrey Peng

Public Playlist

Projects

Jeffrey Peng ยท 2 songs, 5 min 23 sec

#TitleDuration
1

Habit Tracker

habitstracker.ca
  • Built a full-stack habit tracking application using Next.js, TypeScript, Express.js, and PostgreSQL, supporting habit creation, daily check-ins, streak calculation, and goal management.
  • Deployed the system end-to-end on AWS (RDS, Elastic Beanstalk, Amplify), configuring environment variables, API endpoints, and database connections across services; added unit and end-to-end tests with Jest and Cypress
February 2026
2

Food Delivery Time Predictor

  • Performed exploratory data analysis and feature engineering using scikit-learn, pandas, numpy, and matplotlib, applying one-hot encoding, ordinal encoding, and MinMax scaling to normalize features.
  • Created and compared a neural network model using TensorFlow with Keras and a gradient boosting model using XGBoost, tuning hyperparameters with GridSearchCV on validation data to improve accuracy by 8%.
January 2024
3

Spotify Song Recommender

  • Analyzed audio feature data for over 1 million songs and used Spotify's API to extract song features.
  • Scaled, one-hot encoded, and mapped the features of the data in Jupyter Notebook using scikit-learn, pandas, and numpy, and used cosine similarity to measure likeness between two songs.
December 2023