Selected work relevant to Data Analyst / BI Analyst roles and adjacent analytics, automation, and data workflow opportunities. Some projects are internal/private and shown as high-level case studies (no sensitive details or fake links). Links/screenshots can be added later.
Case Study (Internal/Private)
Ontario Public Service - Azure Resource Management App & BI Reporting
Problem: Ministry staff needed a more efficient way to manage Azure resources/VMs and maintain usable reporting.
What I built: A self-service internal app and automated request lifecycle, plus Power BI dashboards/reports and legacy PBIX-to-PBIP modernization.
Outcome: Adopted by staff; supported efficient resource management and saved thousands of dollars monthly on unutilized resources. Also delivered a related wildfire spread prediction POC for decision support.
Power Apps (Canvas/Model-driven), Dataverse, Power Pages, Power Automate, Power BI, PBIX/PBIP, Azure
Case Study (Internal/Private)
GTA Housing Insights Map
Problem: Make housing trend and renter-experience insights easier to explore for stakeholders and renters in the GTA.
What I built: An interactive, filterable GIS map and analysis workflow based on researched/sampled data.
Outcome: Provided filterable insights into housing trends and areas of concern/potential solutions.
QGIS, Data Analysis, Sampling/Statistics, Geospatial Visualization
Project (Open Source, Jan 2026 - Apr 2026)
Service Booking and Consultation Platform
Problem: Streamline service booking/consultation interactions and customer inquiries.
What I built: A full-stack booking/consultation platform with an AI chatbot and supporting documentation.
Outcome: Delivered an end-to-end platform prototype integrating backend, frontend, and LLM-based inquiry handling (repo/demo link to be added).
Python, FastAPI, React, LLM Integration, REST APIs, Documentation
Case Study (Internal/Private)
Ceekam Audit Automation & Reporting
Problem: Manual auditing and transaction tracking were time-consuming.
What I built: Automated Excel workflows and structured transaction/expense tracking and quotation support.
Outcome: Reduced auditing time by 40%.
Excel (Advanced), VBA, Data Cleaning, Reconciliation, Reporting
Project
Trading Bot (Signal Automation)
Problem: Manual monitoring and execution of trading signals across sources is slow and error-prone.
What I built: An automation script/bot that scrapes signals, filters by authorized sources, and executes transactions via trading platform APIs.
Outcome: Demonstrated real-time data processing and API-driven automation workflow (no financial return claims; repo/demo link to be added).
Python, REST APIs, Automation Scripts, Real-time Data Processing
Project (Home Lab, prototype Jan 2025, main Jun 2025, ongoing maintenance)
Home Server & NAS (Proxmox Lab)
Problem: Need reliable self-hosted storage, apps, and hardware-accelerated workloads in a controlled environment.
What I built: A Proxmox VE environment running LXCs/VMs with storage, self-hosted applications, and GPU passthrough for accelerated workloads.
Outcome: Hands-on infrastructure, Linux, networking, and operations experience supporting technical problem-solving.
Proxmox VE, LXC/KVM, Linux, Docker, NVIDIA GPU Passthrough/CUDA, Networking/NAS
CAMLDS Project
Telco Customer Churn Analysis & Streamlit App
Problem: Understand churn drivers and support churn-risk prediction for customer retention scenarios.
What I built: A churn analysis notebook and Streamlit application as part of CAMLDS coursework (Centre of Applied Machine Learning and Data Science, Oct 2026 - Nov 2026, hybrid).
Outcome: Delivered an end-to-end churn analysis and app workflow using Python and machine learning/data analysis tooling.
Python, pandas, NumPy, scikit-learn, matplotlib, seaborn, Streamlit
CAMLDS Projects
Classification Apps & ML Pipelines (Bank Marketing, Employee Attrition, Student Performance)
Problem: Build practical prediction workflows for subscription likelihood, employee attrition, and student pass outcomes.
What I built: Notebooks, preprocessing/model pipelines, evaluation artifacts, and Streamlit apps for each use case (including model comparisons such as Logistic Regression and Random Forest) as part of CAMLDS coursework.
Outcome: Completed end-to-end applied ML projects with saved pipelines/metrics and interactive app interfaces (details available in project files).
Python, pandas, scikit-learn, joblib, Streamlit, Model Evaluation
CAMLDS / Capstone Project
Merged Games Hit Prediction & Analysis
Problem: Analyze and model video game sales/title data to predict hit potential without leakage from outcome variables.
What I built: Data cleaning/merging and EDA across game datasets, feature preparation, hit classification (target: Global Sales >= 1.0M), model pipeline, metrics, and a Streamlit app as part of CAMLDS coursework/capstone work.
Outcome: Delivered an end-to-end analysis and prediction workflow on merged games data with saved pipeline/metrics and app.
Python, pandas, NumPy, scikit-learn, EDA, Streamlit, joblib