I am Steve Harris, a machine learning practitioner and causal modelling researcher from Toronto, Canada. I develop production AI projects by day, teaches analytics at college, and am hopefully heading into a PhD program in the near future. Basically, I can’t stop moving between the lab and the real world.
My professional life is spent at the intersection of data, decision-making, and the systems that connect them. Currently, I work with organizations building recommendation systems, attribution modelling, mixed media modelling, predictive analytics, and multivariate testing with new or existing data sets. I also teach analytics at George Brown College, where I’ve spent over a decade helping students bridge the gap between theory and practice.
My research background is in machine learning and natural language processing applied to data mining, anomaly, and signal detection. My MSc thesis at Athabasca University produced an NLP system for detecting student difficulty in online learning environments. That work, along with a handful of peer-reviewed publications in IEEE and Springer venues that I have contributed to, has taught me something important: prediction is not the same as understanding.
What This Site Is About
This site is where I think out loud about causal inference and causal modelling, including the ideas, the math, the code, and most importantly how to look at a problem and ask the right causal questions.
Most of the ML and analytics world is built on correlations: what patterns exist in the data we have? That’s powerful, but it has limits. It’s the difference between predicting the future by understanding the processes that generate outcomes versus just having the history and hoping it repeats itself.
Causal thinking changes the questions you ask. Instead of what happened, you ask why, or what caused it to happen. Instead of what does the data show, you ask what would happen if we intervened. That shift matters enormously in just about every field that uses data to make decisions.
Here you’ll find posts that walk through how to think about real problems in a causal way, some of the mathematics and theory behind causal inference, and links to my GitHub where I build tools and work through implementations. I also write about the books, papers, and ideas of others that have shaped how I think – not as reviews, but as springboards for working through the concepts myself.
Get In Touch
I’m always happy to hear from people thinking about causal inference, machine learning, or the places where research meets practice. You can reach me at steve@jbs.ca or find me on LinkedIn.