Real Projects. Documented Outcomes.

All Case Studies

Data Engineering & AnalyticsGovernment / Regulatory / Media
Canadian Content Compliance Platform
Canadian Radio-television and Telecommunications Commission (CRTC) — via Oproma Inc.
Challenge
The CRTC needed a system to consolidate, normalize, and validate metadata for Canadian Content (CanCon) compliance tracking across broadcasters. Manual workflows were error-prone and couldn't scale.
50%
Improvement in data insertion speed after batch SQL processing and in-memory caching
50,000+
Weekly records processed through automated compliance checking
What Was Built
• Full ETL pipeline in C# and Python ingesting from CBC databases, Quansic (ISNI/IPN), and MusicBrainz API
• Fuzzy matching with LINQ for deduplication across artist aliases
• MAPL compliance logic engine calculating Canadian Content flags for tracks, contributors, and performances
• Web application with CRUD operations, real-time compliance monitoring, audit logging, and broadcaster reports
Tech
• C#
• .NET
• LINQ
• Python
• SQL Server
• MusicBrainz API
• Quansic API
• MVC
• JavaScript
Data Engineering & AnalyticsGovernment / Enterprise IT
Federal Government Mailbox Migration
Federal government agency (name confidential)
Challenge
Migrate complete mailbox data from Exchange Server 2019 (on-premises) to Exchange Online with zero data loss.
Zero
Data loss — verified through post-migration cross-referencing
Complete
Migration of emails, attachments, calendar entries, contacts, and folder hierarchies
What Was Built
• 7-step ETL pipeline: PST extraction → parsing → EML transformation → validation → Azure AD OAuth2 → bulk upload via Microsoft Graph API and EWS API → post-migration reporting
• Validation scripts checking mandatory fields, attachment integrity, and folder structure
• Error logging and automated retry logic
• Full stakeholder migration report
Tech
• C#
• Microsoft Graph API
• EWS API
• Azure AD
• OAuth2
• Exchange PowerShell
Data Engineering & AnalyticsGovernment / Archives
Document Digitization Pipeline
Library and Archives Canada (LAC) — federal government
Challenge
Process and extract structured metadata from ~1,000 scanned archival PDFs as a proof of concept for scaling to 80,000+ documents.
95%
Automation of document processing workflow
90%
Reduction in OCR errors
30%
Improvement in metadata accuracy
80,000+
Documents validated for pipeline scale
What Was Built
• OCR pipeline using Tesseract with page filtering, rotation detection and correction, and gibberish removal
• Metadata extraction for title, publisher, year, language, authors, edition, and ISMN
• API enrichment and validation via ISBNdb and Google Books
• Automated spell-checking for OCR errors and CSV export for archival systems
Tech
• Python
• Groovy
• Tesseract OCR
• ISBNdb API
• Google Books API
• MiCore
• CSV
Web Application DevelopmentGovernment / SaaS / Document Management
Government SaaS Document Platform (PriviDox)
Oproma Inc. — serving Public Services and Procurement Canada (PSPC)
Challenge
Build and maintain a production document management platform for federal government workspaces — including file management, member permissions, and bilingual reporting — over five years of continuous delivery.
67
Active Government of Canada workspaces
5 years
Continuous production delivery
120+
REST API endpoints
What Was Built
• File manager with tree view, drag-and-drop, versioning, upload progress, and memoization
• Member management with RBAC, role-based permissions, pagination, and bulk operations
• Full reports suite covering workspace activity, team history, disk usage, and permissions
• English/French bilingual throughout; security via 2FA, audit logging, and user action tracking
Tech
• React
• TypeScript
• Redux
• Node.js
• PostgreSQL
• Nginx
• REST APIs
• GitLab CI/CD
• Docker
Data Engineering & AnalyticsFood & Beverage / Small Business
Restaurant Sales Analytics
Hadi's Shawarma (local restaurant)
Challenge
3,000+ transactions across DoorDash, Uber Eats, and SkipTheDishes with no analytical insight. The owner wanted to understand customer behaviour and increase off-peak sales.
+15%
Increase in off-peak sales in the first month after implementing recommendations
3,000+
Transactions analyzed across three delivery platforms
What Was Built
• Raw data ingested from three delivery platforms into Azure Blob Storage
• ETL pipeline using Azure Databricks and Apache Spark to clean, standardize, and deduplicate
• Sales trend analysis: peak order times, order frequency, customer preferences
• Off-peak underperformance identification and targeted promotion recommendations
Tech
• Azure Blob Storage
• Azure Databricks
• Apache Spark
• Python
• Matplotlib
• Power BI
Data Engineering & AnalyticsFood & Beverage / Analytics
Sales Dashboard — 1 Million Transactions
Balaji Foods (simulated data, real methodology)
Challenge
Build a structured ETL pipeline and Power BI dashboard to analyze 1,000,000 transactions — covering product performance, time-based sales patterns, and payment type breakdowns.
1,000,000
Transactions processed through the ETL pipeline
$275K
Total sales revenue analyzed
1,361
Units — top volume item (Cold coffee)
What Was Built
• Python and Pandas ETL pipeline: raw CSV ingestion, date parsing across inconsistent formats, validation, deduplication, and feature engineering
• Power BI dashboard with DAX measures for product performance, weekday/time-of-day heatmap, and payment analysis
• Stakeholder-ready insights for inventory planning and staffing optimization
Tech
• Python
• Pandas
• Power BI
• DAX
• CSV
Web Application DevelopmentWeb Application / Sports Analytics
FantasyPLStats — Fantasy Premier League Statistics App
Personal project
Challenge
Build a real-time statistics platform for Fantasy Premier League data with machine learning insights — including automated daily updates and a sortable, filterable frontend.
160,000+
Records in the ML training and analytics pipeline
What Was Built
• Real-time data pipeline from FPL API to DataStax AstraDB with Azure Functions cron job for automated daily updates
• ML pipeline in Azure ML: correlation analysis, feature selection, and model training
• Next.js frontend with React Table for sorting, filtering, and player comparison
• Azure Static Web Apps deployment with Azure DevOps CI/CD
Tech
• React
• TypeScript
• Next.js
• Node.js
• DataStax AstraDB
• Azure ML
• Azure Functions
• Azure Static Web Apps
• Azure DevOps