A specialist approach to data
MG Data Analytics is an independent data analytics practice based in North York, Ontario, working remotely with clients across the province.
The practice is built on a PhD in Statistics and over seven years of applied data work. Many providers can build a dashboard. Fewer can tell you whether the underlying model is sound, whether your sample supports the conclusion you want to draw, or which method actually fits your data. That distinction is where this practice operates.
Selected work
Donor segmentation and retention modeling - nonprofit sector
Segmented 10,000+ donor records using RFM feature engineering and K-means clustering, then built ensemble models to predict donation recurrence at 82% accuracy. Delivered a Power BI dashboard for stakeholder review, with segment findings used to shape retention strategy.
Statistical process control - manufacturing
Applied control charts and variance decomposition to isolate sources of process variation on a production line, and developed predictive models to improve stability. Results included an 8% reduction in defects and a 5% reduction in output variability.
Large-scale survey design and analysis - public sector
Led six survey projects covering 10,000+ households to develop economic accounts supporting government policy planning. Managed validation and analysis of 40,000+ data points, improving collection efficiency by 15% through instrument and process refinements.
An independent data analytics practice working with organizations that need more than descriptive reporting.
Small and mid-sized organizations across Ontario, manufacturers, and research teams needing analysis they can defend.
We match the method to the question, validate the result, and explain it in language you can act on.
Statistical depth, not just reporting
Most analytics work stops at describing what happened. A PhD-level statistical background means the methods are chosen deliberately and the conclusions hold up when questioned — whether that's model validation, sampling assumptions, or knowing when a result isn't significant.
Direct access to the analyst
You work with the person doing the analysis. No account managers, no handoffs, no junior staff learning on your project.
Results you can act on
Findings are delivered in plain language with clear recommendations, not a technical report you need someone else to interpret.