Projects
Data science and engineering work I’ve built or shipped — mostly prediction models and the pipelines that feed them.
Water Main Break Prediction
The Water Main Break Prediction is a web application which predicts the probability of water main breaks for the municipality of Kitchener-Waterloo. It is made using Streamlit and deployed on Heroku.
Instacart Market Basket Analysis
This project applies Market Basket Analysis on the Instacart Online Grocery Shopping dataset using Databricks, Spark, and Spark SQL, ingests data through an ETL pipeline, performs exploratory analysis and employs machine learning with PySpark and Scala to discover frequent itemsets, informing purchase recommendations.
DoorDash Delivery Duration Prediction
The aim of this project is to predict the delivery duration of DoorDash orders. The model is trained on deliveries made in 2015 and utilizes classical regression techniques as well as boosted tree regressors.
Lyft Driver LTV Prediction
The goal of this project is to predict the lifetime value (LTV) of a driver for Lyft and identify the main factors affecting a driver's LTV. I also aim to explore whether there are specific segments of drivers that generate more value for Lyft than the average driver, and provide actionable recommendations for the business based on the findings.
Retail Sales Analysis
This project aimed to analyze transaction data for chip sales in order to understand customer behavior and optimize store strategy. The data consisted of customer transaction records from a national grocery store chain over a 12-month period.