Selected technical work

Data projects that make decisions clearer.

A focused portfolio across data engineering, database development, analytics, and decision systems - from raw data collection to decision-ready dashboards.

Data Engineering Database Development Data Analytics Business Intelligence

Project archive

Selected case studies

04 / 04 published

PROJECT_01 / MARKET_INTELLIGENCE

Toronto African & Caribbean Restaurant Market Intelligence

A market intelligence workflow that turns customer reviews into a practical launch blueprint for an African and Caribbean fusion restaurant. The project brings together web-sourced review data, a PostgreSQL data model, cloud NLP sentiment analysis, keyword extraction, and an interactive reporting layer.

Python Pandas PostgreSQL SQLAlchemy Google Cloud NLP TF-IDF Looker Studio
Open Looker Studio dashboard Open project notebook Open in Google Colab

Business question

What does the Toronto market say about a new fusion restaurant?

Review-level analysis was used to identify customer sentiment, recurring themes, and service or menu signals that can inform a market-entry plan.

Database work

From flat reviews to an analysis-ready relational model.

  • Restaurant, review, and analysis tables
  • Cleaned, deduplicated review records
  • SQLAlchemy-based loading to PostgreSQL

Analytics output

Operational signals, not just charts.

  • Mean sentiment score: 0.497
  • 67% of reviews expressed satisfaction
  • Cuisine-segment comparison
  • Top customer keywords and themes

PROJECT_02 / ECONOMIC_INTELLIGENCE

Economic Intelligence Data Platform

A production-style Azure data engineering platform that ingests, transforms, and serves Canadian and G7 economic indicators for analysis. It brings together multi-source orchestration, a bronze-to-gold lakehouse design, analytics-ready SQL views, and infrastructure managed as code.

Azure Data Factory ADLS Gen2 Azure Synapse Delta Lake Terraform Python SQL Power BI
Explore GitHub repository

Data engineering

One orchestrated platform for four public data sources.

Source-specific ADF pipelines feed a master pipeline that runs ingestion in parallel and lands raw files in the bronze ADLS Gen2 zone.

Architecture

Medallion design built for dependable analytics.

  • Bronze: raw ZIP and JSON source files
  • Silver: cleaned, structured Delta tables
  • Gold: curated views for dashboards and analysis

Evidence in the repo

Implemented pipeline, transformation, and serving layers.

  • pl_master_ingestion.json orchestration
  • 05_silver_to_gold.py transformation
  • vw_unemployment_trend.sql analytics view
  • Terraform environments and DQ workflows

PROJECT_03 / CALL_CENTER_ANALYTICS

Dialer Performance & Call Center Analytics in SAS

An end-to-end SAS Studio analytics solution that transforms PwC Call Centre Trends data into a dialer-style operational model. The work turns 5,000 call records into management-ready KPIs for call outcomes, agent performance, dialing windows, speed of answer, and customer satisfaction.

SAS Studio Base SAS DATA Step PROC SQL PROC SGPLOT ODS Graphics CSV
View SAS source Open results report

Operational baseline

5,000 calls modeled into actionable service KPIs.

  • 81.1% answer rate (4,054 answered calls)
  • 18.9% abandonment rate (946 calls)
  • 89.9% resolution rate among answered calls

Performance signals

Identify where operations can improve service delivery.

  • Streaming had the strongest answer rate: 82.9%
  • Technical Support had the lowest answer rate: 79.0%
  • Waits over 120 seconds had the lowest satisfaction: 3.28

SAS deliverables

Reusable KPI tables and management-ready visualizations.

  • Answer, abandonment, and resolution by topic
  • Agent resolution and satisfaction comparisons
  • Hourly and weekday dialing-window analysis
  • Speed-of-answer satisfaction trend analysis

PROJECT_04 / NETFLIX_CONTENT_STRATEGY

Netflix Content Strategy Analysis

A marketing analytics and data engineering project that turns a messy Netflix content catalogue into competitive intelligence for a Canadian streaming platform. The pipeline cleans and enriches content metadata, loads it into a normalized PostgreSQL model, and surfaces portfolio, regional, and acquisition insights in Power BI.

Python Pandas NumPy Google Gemini PostgreSQL ETL Power BI DAX
Read project report Explore GitHub repository

Catalogue profile

8,797 titles reveal a movie-led portfolio.

  • 69.69% movies and 30.31% TV shows
  • 6,131 movies and 2,666 TV shows
  • International Movies, Dramas, and Comedies lead genres

Market intelligence

Geographic and timing patterns inform expansion decisions.

  • United States accounts for 56.48% of titles
  • North America represents 41.63% of catalogue content
  • 63.14% of titles were released within the previous two years

Data & BI delivery

From unstructured data to an analytics-ready platform.

  • Metadata enrichment and validation
  • Relational PostgreSQL integrity and indexing
  • Power BI model, DAX metrics, and executive dashboards
  • Clear scope: catalogue strategy, not viewing behavior

More projects in progress. This page will continue to grow with data engineering, database, and analytics case studies.

ARCHIVE_ACTIVE