Like trees in my mind
Senses fractal out of bounds;
Each branch has a branch
Along which world flows into me
Leaving traces, data signs …

It is 4am. My daughter is seizing. I am carrying her as she flails unconsciously, yelling to my wife to get her rescue meds, grabbing the iPad to get footage for the neurologist …
I am observing my daughter’s convulsions simultaneously as her panicked father and with another aspect of my persona, cooly recording micro details I will write into a seizure journal before I forget them with the passing wave of adrenalin. Is her colour good? Is she flailing equally on both sides? Are her eyebrows fluttering? Is she choking? In the seconds before my wife joins me to administer the rescue meds I am in a dark hallway holding a seizing child and my mind is afire with hypothesis, inferences, observation — the stuff of data science, the stuff of science.

I have spent much of my adult life lost in data, navigating data points as if they were new continents. In my day job I am a data scientist and analytics architect. But what I do there is ultimately not so different than what I am doing here while holding a convulsing child. I am rapidly making inferences with as much information I have at hand — always wanting more — and trying to determine a course of action consistent with the information. Then, with reflection, I am capturing insight into systems of code; building architecture. If analysis is the story in the data, architecture is the city in which a million stories play out.

In my day job I build inference engines, analytic software systems, simulations, predictive models. Holding my daughter in my hands, damping down the terror she is dying, I have only my knowledge of complex networks, simple models of her brain and my imagination as stand ins for inference engines. Similarly, Julie has her knowledge as an occupational therapist, her working understanding of brain anatomy. We share training in classical and population genetics. Out of need and desperation we become not only husband and wife — but a research team, seeking a model, that can guide our observations into therapy. We build a cartoon model in words — loose analogies to the mathematics and algorithms that will be needed to flesh the model out. Like all models, it is incomplete — but it gives us a frame to hang our decisions upon. It gives us a basis for action. It allows us to anticipate and be proactive. It helps us keep our daughter alive.

The topic of this blog is the thought process behind data science, the ‘psychology of data science’. Using historical and current examples from science, medicine and industry as well as stream-of-Python illustrations of analytical exploration in progress I hope to illuminate the unique mix of insight, science, detective work, and art that grounds data science. Data science being a current and popular term for a very old practice: understanding our world from data, anticipating the future and acting in the present to realize that future. Whether that future concerns a forest, a community, a business, or a playful child with a sodium channel mis-sense mutation causing epilepsy.