Understanding Data Science with GCSE Maths - Embracing the Black Box

Richard Pike, Head of Client Services at Profusion

I gave up maths as soon as the education system allowed me. It was a moment of joy, wonder and relief. I then quite deliberately proceeded to spend the next 20 years having as little to do with maths as possible, as I made my way in the fluffy world of Humanities and Creative Advertising.

I’m now a Data Consultant and have been for over 18 months now.

So how on earth can I do my job when I wouldn’t recognise a Bayesian Regression if it tried to buy me a drink in a bar? It’s a question I have got (suspiciously) often from ex-colleagues, friends and family.

It was something I thought about very carefully before taking the plunge from fluffy advertising into hard data, not least because I was unlikely to get through an interview without being able to answer this question. Once I actually started the role, I discovered all the positive intentions and self-directed learning could not even scratch the surface of what Data Science is, does and can do. ?

After the first weeks of getting lost in technical explanations of how certain Data Science concepts work, I was close to giving up. It’s a hard road, when even Wikipedia is incomprehensible.

For the first time in my life, I literally couldn’t get my head round concepts and ideas.

All I could do is go back to what I did know – words. Rummaging around in barely remembered memory crypts back from when we were literally partying like it was 1999 or (more realistically) seeing something on Instagram, I found ignorance. Specifically Socratic ignorance and “the only thing I know is I know nothing”. Instead of trying to know, I embraced not knowing.

In short I embraced the Black Box.

Data Science built a model that does stuff, how it does it I was clueless – it was a black, unfathomable box. But I did know what goes in and what comes out. Focusing on the input and output meant that I could use the expertise I’d been hired for – understanding marketing and our client’s business. So instead of worry about how the Black Box worked, I focused on making sure what we got out of the black box was the right answer.

The greatest barrier to understanding Data Science it turns out is linguistic not mathematical. Two Towers, Random Forest, Skynet, Bayesian Regression, all of these sound mysterious, faintly terrifying and very complex. Even the word Algorithm is bafflingly difficult to spell (surely it has a y in it somewhere)??

I simply hadn’t been asking the right questions – I had been focusing on “how” in an eternally doomed attempt to teach myself advanced statistics. Once I started to ask “why?” it opened up everything.

By asking why, you can understand the decisions were made and interrogate those. In doing this, you can begin to not only understand the decisions but also a little of what’s in the mystery black box…?

This is why Profusion’s Data Literacy training courses focus on asking the right questions, equipping businesses with the ability to ask why. If you’d like to understand more about Data for Leaders, get in touch with the experts: [email protected] or hit me up on LinkedIn.?

Richard Pike, Head of Client Services at Profusion