Projects & Systems
Platforms, Models, and Tools Built to Solve Real Problems
My work has ranged from enterprise data platforms and decision-support systems to predictive models, geospatial analytics, and custom applications. Some projects reduced months of manual effort to minutes. Others saved millions of dollars, improved governance, or gave leaders access to information they had never been able to trust before.
What connects them is practical value: useful systems, explainable logic, and solutions designed to hold up in the real world.
A Portfolio of Practical Systems
I am at my best when a problem is messy, important, and not yet well solved. Across geospatial analysis, data science, and modern data platforms, I have consistently gravitated toward work that combines technical depth with operational usefulness. In some cases that meant building governed cloud-based data products. In others, it meant designing predictive models that could survive regulatory scrutiny, or building spatial analytics that changed how high-value assets were evaluated.
The projects below reflect that range. They are not just examples of things I have worked on. They are examples of how I think: understand the real problem, build the right solution, and make sure the result can actually be trusted and used.
Modernization, Planning, and Trusted Delivery
HR Lakehouse Environment
Built a modern Databricks-based HR data environment from the ground up, including direct API integrations to Workday and SAP, governed medallion architecture, controlled schema access, automated pipelines, and predictive models supporting scorecard metrics and workforce analytics.
Workforce Planning Transformation
Reworked a vendor-delivered workforce planning solution that had become fragile, complex, and time consuming to maintain. The existing process relied on disconnected tools and manual upkeep that consumed weeks each month, limiting planners’ ability to focus on strategic work.
Compensation & Long Range Planning
Designed a position-level compensation forecasting model combining workforce plans with wage, bonus, benefit, pension, allowance, vacancy, and currency data. The process moved from a difficult, time-intensive workflow to a repeatable on-demand capability that could produce forecasts in about an hour before validation.
People Metrics Platform
Migrated HR reporting from Excel-based Workday extracts and manual calculations to direct Workday API ingestion and Databricks-based metric calculation, delivered through PowerBI. Expanded from a small number of manual metrics for one team to more than one hundred standardized metrics across dozens of teams.
CRUD-Style HR Applications
Designed and built user-driven HR applications that extended beyond static reporting and into transactional workflow support, helping HR functions move from passive information delivery toward practical operational tools.
HR Data Strategy & Governance
Helped define the broader HR data strategy, governance model, and platform modernization approach, moving the organization away from fragile spreadsheets and manual processes toward governed, trusted data products.
Production Models with Real Business Value
Electricity Demand Peak Prediction
Built a probabilistic electricity demand model to predict Alberta coincident peak periods, where poorly timed cogeneration outages could cost millions. Historical demand, weather, and time-of-day effects were used to train thousands of neural networks in TensorFlow and PyTorch, with results delivered through PowerBI. The system has continued to provide value over multiple years and helped avoid several million dollars during at least one maintenance event alone.
Emissions Model / PEMS
Developed a predictive emissions monitoring system as a potential alternative to costly continuous emissions monitoring hardware. In collaboration with engineering and the regulator, more than 200 candidate process tags were evaluated, models were iterated into the hundreds, and performance exceeded required thresholds. The result became the first regulator-approved PEMS model in Western Canada, establishing a foundation for millions in potential annual savings.
Production AI Systems
Built and deployed machine learning systems into production environments where explainability, durability, and scrutiny mattered. This work emphasized practical governance and long-term usability, not just model performance in isolation.
Monte Carlo & Scenario Simulation
Built simulation-based tools to support uncertainty modelling, planning, and long-range decision support, helping teams reason about risk and future scenarios rather than relying on single-point assumptions.
Model Governance in Practice
A recurring theme in my modelling work has been knowing when not to force deployment. In the emissions project, drift risk was identified during ongoing monitoring and the decision was made to suspend deployment despite remaining within regulatory bounds, prioritizing environmental protection and sound governance.
Company-Wide Enablement in R & Python
Trained staff across the company in R and Python programming, helping build data literacy and reduce reliance on a small number of specialists. This enabled broader use of modern analytical methods beyond the data science team itself.
Spatial Intelligence, Asset Risk, and High-Value Decision Support
Lease Exposure / MLE
Transformed a largely manual lease evaluation process tied to Alberta oil sands Minimum Level of Evaluation obligations. What had taken roughly nine months each year through map measurement and manual well checking was reduced first to about a week through geospatial analysis, then to minutes at provincial scale using Python and vendor GeoLogic data. The work surfaced hundreds of millions of dollars in potential liability and improved proactive planning and budgeting.
Heritage Royalty Divestiture Support
Led the data component of a high-pressure, highly confidential royalty divestiture following the acquisition of ConocoPhillips’ Canadian assets. Built a single authoritative dataset by reconciling mineral land, drilling, exploration, accounting, and production data under aggressive timelines. The divestiture closed as the largest Canadian E&P royalty transaction in history, generating approximately $3.3B in proceeds.
GeoPlanNet / Patented Innovation
Helped adapt idealized network optimization concepts into a probabilistic GIS-based planning method that accounted for reservoir uncertainty, environmental constraints, and phased development realities. The resulting GeoPlanNet methodology supported more environmentally conscious and capital-efficient infrastructure planning, was contributed to COSIA, and led to a patent filing.
Acquisitions & Divestitures Support
Provided spatial and non-spatial analytics support for multiple confidential acquisition and divestiture initiatives, including valuation support, risk assessment, transaction readiness, competitive intelligence, and bid strategy.
Enterprise Spatial Data & Vendor Design
Maintained enterprise spatial datasets and worked closely with external vendors to design custom geospatial data products needed for advanced analysis. Several major projects depended on restructuring vendor data into forms suitable for scalable internal use.
Geospatial Services Leadership
Managed and helped lead geospatial analysts supporting land, environment and regulatory, ground deformation, investor relations, communications, acquisitions and divestitures, and other corporate functions, combining delivery leadership with hands-on analytical work.
Side Projects That Still Reflect How I Work
NHL Analytics & Draft Application
Built a Python and Flask application to run a custom hockey pool draft, including user login, draft mechanics, nightly API ingestion, database-backed storage, analytics, and live dashboards. It started as a hobby project and became a practical full-stack exercise in data products, automation, and UX.
Photo Organization AI
Built tools to scan, group, and organize large personal photo collections using metadata, file discovery, and machine learning concepts to surface duplicates and related images.
Data Experiments & Learning Builds
I regularly build smaller projects to learn new tools, frameworks, and patterns, usually by applying them to real problems I care about. These experiments are a meaningful part of how I stay current and deepen my technical range.
The projects I value most are the ones that solve a hard problem, hold up under scrutiny, and leave behind a stronger capability than existed before.
Brett Wiens 