Data Science × Process Engineering
StateSpace Analytics blends engineering with data science, applying statistics and physics-based modelling to industrial systems — turning process data into models you can monitor, simulate, and act on.
§ 01 — Capabilities
Six disciplines, matched to the question you need answered — from monitoring a live process to simulating one that doesn’t exist yet.
Multivariate process monitoring, control, and optimization for continuous and batch processes. Detect anomalies, leverage built-in diagnostics to drill down to root causes, reduce variability, and optimize product quality.
Inference that reports what you actually know. Models return full posteriors and credible intervals, carrying uncertainty through to the decision instead of collapsing it into one number.
First-principles and grey-box models of the plant — thermal, fluid, mechanical. Capture the dynamics in simulation and test changes before they touch the real system.
The most information from the fewest runs. Isolate the effects driving your process and locate the settings that matter.
Throughput, queues, and resource contention, modelled event by event. Size capacity, expose bottlenecks, and stress-test operations before committing capital.
Quantify what can go wrong and how likely it is. Reliability models and Monte Carlo propagation turn component failure rates and uncertainty into system-level risk — and rank the contributors worth mitigating first.
§ 02 — Selected work
A sample of engagements — from rolling mills to shipping ports to building environmental systems.
Metals · Rolling
Modelling strip thickness and flatness across a rolling line, then tuning setpoints to hold profile within tolerance — mill physics, designed experiments, and multivariate modeling working together.
Energy · Pipelines
Quantifying leak and failure risk across a pipeline network while optimizing pump scheduling and energy use — fault- and event-tree risk models alongside hydraulic and energy modelling.
Medical devices · Regulated mfg
Process characterization and statistical monitoring for a regulated production line — capability studies, designed experiments, and control charts built to hold up under audit.
Maritime · Logistics
Discrete-event models of port and fleet operations — sizing berth and vessel capacity, exposing bottlenecks, and stress-testing schedules before assets are committed.
Buildings · HVAC
Physics-based models of building ventilation and thermal behaviour — capturing airflow and energy dynamics to test control strategies and equipment changes in simulation.
§ 03 — Approach
A working sequence, not a fixed package — each step earns the next.
Pin down the process, the question, and the measurement that defines a good outcome.
Get the right data and physical understanding in place to answer it.
Build the statistical or physical model — then validate it against reality.
Put the result into operation, documented so your team can run it.
§ 04 — About
Rigorous by default — and built to be handed over.
StateSpace Analytics is a consulting practice for organizations that treat their processes as systems worth measuring. The work sits where statistics meets engineering: enough physical modelling to respect how a plant actually behaves, and enough statistical discipline to know when a result is real.
Every model is built to be audited, re-run, and maintained by your own team — not left behind as a black box.
§ 05 — Contact
Tell us about the system and the decision behind it. We’ll tell you whether the data supports the answer you’re after.
info@statespaceanalytics.ca