Green AI · Sustainability
Coming Soon
How Green is My AI?
Energy audits, data centres, and reducing your carbon footprint.
Portfolio · Hamilton, Ontario
JoAnne
Hosick
Physicist · Data Scientist · AI Builder
I've spent 20+ years at the intersection of scientific rigour and real-world systems — from Ontario's electricity grid to renewable energy analytics to the corner of my kitchen table. I like data that has a lot of questions. Here's some of what I've built.
Selected Projects
Things I've Built
Data Analysis · Smart Home · Live Dashboard
HVAC Energy Analysis
Our heat pump was running its resistance strips when it had no reason to. I built a dataset to prove it.
Collected 16 months of Daikin Skyport cloud data and 6 months of hourly readings from a Sense Energy Monitor, cross-referenced with Open-Meteo hourly weather reanalysis for Hamilton. The air handler consumed 3,492 kWh over 6 months — 37% of combined HVAC energy — despite being a supplemental source that should only activate in extreme cold. The Pearson correlation between air handler wattage and outdoor temperature was −0.15 (R² = 0.02). A properly staged system should show −0.65 to −0.80. Issue identified, documented, and resolved. The live dashboard below tracks post-fix system performance as data accumulates across the next heating season.
Launch Dashboard- Sense Energy Monitor
- Daikin Skyport
- Open-Meteo API
- Pearson Correlation
- Python
- Statistical Analysis
- Live Dashboard
Creative AI · Interactive Experience
Work in Progress · Still a bit glitchy
Universe Engineer
What does it take to simulate a universe — from the first hydrogen atoms to planetary life — one slider at a time? An evolving experiment in creative AI and interactive narrative.
Universe Engineer is a 35-page interactive experience where the reader dials in the laws of physics and watches a procedural cosmos respond. Five sliders control conditions from stellar density to orbital mechanics; particle scenes respond in real time as nebulae collapse, stars ignite, planets form, and (eventually) life appears. A Game/Explore mode chooser and Maximum Drama preset let you be either careful scientist or chaos agent.
Built entirely with Claude as the coding partner across six development phases — from skeleton scaffold through narrative passes, visual polish (bloom, blackbody colour, volumetric nebula, curl-noise fields, N-body perturbations), to photo-seeded particle systems using real cosmic imagery. Video sequences were generated with Seedance. I also ran a Blender procedural render pipeline for the cinematic scenes, but didn’t like where it landed — so those were replaced.
Launch Universe Engineer- Claude (AI coding)
- Seedance (video generation)
- JavaScript / Canvas
- WebGL2
- Particle Systems
- Procedural Animation
- N-body Simulation
Automation · Family Tech
The Morning Briefing
Every morning, my daughter and I get an email that knows what changed since yesterday, what's on today, and who's driving on Saturday.
My daughter has a busy schedule — piano lessons, school events, science programs, pizza days. Every morning, an automation scans our incoming emails to detect new activity notices, updates a shared Google Sheet, diffs it against the previous day's snapshot, and generates a formatted daily briefing: changes detected (events added or removed), what's happening today, and the full week ahead with times, locations, and driver assignments. No manual input required.
It knows when to keep things short and when to flag a busy stretch. It's a small system, but it completely eliminates the mental load of keeping two schedules synchronized across two phones.
Good morning! · Happy Sunday, April 19
SPREADSHEET CHANGES TODAY
✕ Apr 18 – Science program (online session – completed!)
+ May 15 – School pizza day
WEEK AHEAD
Thu Apr 23 Piano Lesson – 6:00 PM (Driver: Jo)
Sat Apr 25 Science program – 10:00 AM – Local museum
Sat Apr 25 Science workshop – 11:00 AM – Online
- Email Scanning
- Google Sheets API
- Python
- Cron Scheduling
- Diff / Change Detection
- Automated Reporting
AI Infrastructure · Enterprise
Enterprise ML Data Lab
I built the environment where analysts became data scientists — and watched the user community grow 7× in a year.
As part of a six-project, $3.5M Enterprise Data Strategy at Ontario's electricity grid operator, I stood up an Azure-based Machine Learning Data Lab and ran the RFP and vendor negotiations for an AI development platform and enterprise data catalogue across 8 departments. The lab accelerated smart-metering analytics 15× on hourly consumption data from 5 million meters.
I built the community around it — training videos, use-case libraries, a cross-disciplinary community of practice. Active users grew from 20 to 135+ analysts and engineers. Neural-network models built in the lab improved real-time generation dispatch accuracy by ~200 MW across Ontario's grid.
- Microsoft Azure
- ML Ops
- Python
- Neural Networks
- Enterprise RFP
- Data Catalogue
- Community of Practice
App Development · Business Automation
3D Print Shop Manager
Running a small 3D printing business means juggling orders, queues, materials, and customer updates all at once. I built an app to handle it.
The Print Shop Manager is a custom React web app designed to automate the operational overhead of a home-based 3D printing business. It tracks incoming orders, manages the print queue with priority scheduling, monitors filament inventory, and calculates per-job costs including material, time, and overhead.
Automated customer notifications go out at key stages — order confirmed, print started, ready for pickup — without any manual action required. The dashboard surfaces what needs attention today, what’s in progress, and what’s waiting, so the operator can focus on producing rather than administrating.
- React
- Order Management
- Queue Scheduling
- Inventory Tracking
- Automated Notifications
- Job Costing
About
Physicist by training.
Physicist by training.
Builder by practice.
I started out simulating galaxy collisions. My M.Sc. at Western Ontario was in Scientific Computing — I was the first graduate of the program — and my thesis involved running N-body simulations on SHARCNET supercomputers. I published three peer-reviewed papers on variable-star photometry before pivoting to everything that followed.
That was 2003. Since then I've spent two decades building the data infrastructure that powers large, complex systems. Fifteen years at Ontario's electricity grid operator (IESO), where I eventually led the Enterprise Data Strategy — a $3.5M program across 8 departments that built an Azure ML stack, a machine learning lab, and data pipelines running on 5 million smart meters. Then Recurrent Energy, where I manage analytics for 55 North American solar and battery sites and lead a $500k cross-functional AI initiative to detect equipment underperformance.
I'm currently completing a Master of Engineering in Interdisciplinary Artificial Intelligence at the University of Ottawa.
The thread through all of it: I like systems that have a lot of data and a lot of open questions. I like finding the answers — whether that's a 200 MW improvement in Ontario's generation dispatch, a $700/year heating bill reduction, or a morning email that knows it's going to be a big Saturday.
Career
Background
-
Now
M.Eng Interdisciplinary Artificial Intelligence University of Ottawa, Faculty of Engineering — In Progress
-
2025
Senior Manager, System Performance & Data Analytics Recurrent Energy — Analytics for 55 North American solar & battery sites (3+ GW)
-
2024
Manager, Analytics, Operations & Maintenance Recurrent Energy — Led $500k AI initiative across 5 departments; 20% reduction in operator workload
-
2019
Senior Manager, Data & Analytics — Smart Metering IESO — $3.5M Enterprise Data Strategy; ML Data Lab 20→135+ users; 15× analytics acceleration on 5M meter dataset
-
2009
Planner / Engineer / Senior Engineer IESO — Neural-network models for DER output; improved real-time market accuracy ~200 MW
-
2006
Engineer Trainee / Dispatcher Hydro One — Transmission Planning, Operating Technology, Ontario Grid Control Centre
Outreach & Research
Outreach
- IESO delegate, Electricity Canada Data Analytics Working Group (2020–2024)
- Speaker, Big Data & Analytics for the Public Sector (2022)
- Host & moderator, IESO Learning Week — cross-sector AI discussion featuring Vector Institute CCO
- STEM outreach: University of Toronto "Math in Motion" for Grade 9 girls
- Technical advisor, Ontario Science Centre Electricity Show
- Public planetarium presentations; assisted installation of a radio telescope
Publications
49+ citations · ResearchGate: J. Hosick
- Percy, Hosick & Leigh — Self-Correlation Analysis of RV Tauri Stars, PASP 2003
- Percy & Hosick — Do Red Giant Stars Pulsate in High Overtones?, MNRAS 2002
- Percy, Hosick, Kincaide & Pang — Autocorrelation Analysis of Short-Period Be Stars, PASP 2002
Get in Touch
Let's talk
Let's talk
data.
I'm happy to talk about AI, the electricity grid, variable stars, or automating things that shouldn't need to be manual.
joanne@quantumboutique.caHamilton, Ontario