back&forth @ Learning Analytics Hackathon 2024
October 2024
- Engineered a learning analytics tool using the Canvas API, Python, and R to extract and analyze online course discussions, explore participation trends, resulting in data-driven course improvement recommendations.
- Performed HTML data cleaning with BeautifulSoup on 70+ discussion threads. Structured the content into CSV format, complemented by sentiment analysis using TextBlob to identify student interaction patterns.
- Introduced back&forth, an AI chatbot designed to encourage deeper engagement among students by prompting thoughtful interactions and providing personalized feedback.
Uncovered Interest Rate Parity: Empirical Study with Linear Regression
August 2024 – Present
- Performed an in-depth empirical study on uncovered interest rate parity (UIP) using linear regression models to assess if UIP holds in practice.
- Analyzed secured and unsecured risk-free rates such as SOFR, CORRA, SONIA, TONAR, sourced from reputable financial institutions including the Bank of Japan and Federal Reserve.
- Gained proficiency in using a Bloomberg Terminal for real-time data gathering and economic research, which significantly enriched my analysis and understanding of global interest rates.
Portfolio Website (danialramzan.github.io)
August 2024 – Present
- Built a minimalist, responsive website using Jekyll, HTML, CSS, and Liquid, with deployment managed through GitHub Pages and CI/CD.
- Configured local development environment using Ruby, leveraging Gemfiles for dependency management.
- Employed Google Analytics for detailed insights into visitor behavior, enabling data-driven improvements.
- Ensured a responsive design for a consistent & intuitive user experience across all devices (tested on 5+ devices).
Dynamic Billiards Simulator: Physics Simulation Project
July 2024 – Present
- Simulated complex billiard collisions in Python, incorporating both linear and angular dynamics with a variable coefficient of restitution (COR) to model a wide array of realistic physical interactions.
- Enhanced the simulation’s complexity by introducing features to approximate the value of pi to 7+ decimal places through dynamic collision modeling, based off a mathematical paper.
- Planned future developments include a GUI implementation, visual physics engine, and porting the program to C++ for improved computational efficiency.
Diabetes Classifier: Logistic Regression Project
July 2024 – August 2024
- Conducted a data science project predicting diabetes risk using the Pima Indians Diabetes dataset, assessing model limitations and accuracy.
- Collaborated with peers on exploratory data analysis (EDA), data preprocessing, model training, and model evaluation, achieving more than 80% accuracy on test data.
- Documented model challenges and proposed improvements through alternative algorithms and data handling techniques for better prediction accuracy.
Social Media Scraper: Automated Data Extraction Tool
January 2024
- Constructed a Python-based data scraping tool using BeautifulSoup (BS4) and Selenium to gather market rent prices to help price my summer sublet.
- Extracted post data from social media groups by interpreting HTML class names and handling errors.
- Implemented data cleaning and transformation processes using Pandas to support downstream analysis, storing results in CSV format for easy access.
Breaking Bad Habits Android App @ HackCamp 2023
November 2023
- Developed Android app in a team of 4 to address substance abuse in British Columbia using Android Studio (Java).
- Created a logging system for users to track progress, emotions, and challenges in their rehabilitation journey.
- Incentivized self-care tasks by adding an interactive map of nearby mental health facilities with ratings.
FinanciallyFit: Reverse Attendance-Based Java Billing App
September 2023 – December 2023
- Pioneered a novel full-stack gym billing system, which imposes financial penalties for missed gym sessions.
- Implemented user management, bill generation, data persistence, etc., with testing using JUnit and an intuitive GUI.
- Achieved 102% on the course project, integrated user stories and CI/CD principles, streamlining future add-ons.
Lead Data Scientist, Pulsar Classifier
March 2023 – April 2023
- Led a team of 4 on a project to accurately classify pulsars using the KNN machine learning classification algorithm on the HTRU2 dataset, focusing on 8 key variables for predictive analytics using R.
- Produced a pulsar classifier with a >90% accuracy on test data while avoiding overfitting. Achieved 95% in the project, performing EDA, model training, and outlining our process in Jupyter Notebooks for a streamlined workflow.