Most people, when they hear the words ‘data science’, imagine a field that draws entirely on the far left side of the brain. A world of input and output, algorithms and calculation that provides you with the hard analysis you need to make objective decisions.
It’s a well-intentioned but misleading perspective.
The truth is, data science is multifaceted. It’s investigative work, it’s innovative, and it’s an evolving field - one where technology, methodology and problem solving collide.
And fundamentally, at its heart, data science is a creative process.
Data science is the art of helping humans make better decisions
Contrary to popular belief, data science isn’t about algorithms making decisions. It’s not about handing technology the steering wheel and telling it to drive where it will. Data science is for humans, not machines. You could define it as ‘the art of helping humans make better decisions.’
Most people would include the word ‘data’ in a definition of data science. But don’t be misled into thinking that data science is primarily about the data. It’s actually about people.
At Unai, much of our focus is on extracting the tasks that are of no value for people to do manually but take huge effort. Our tools don’t make crucial decisions, they simply automate the most mundane parts of the process so people can focus their human intelligence where it’s needed.
You can see this in the algorithms we created for the University of Exeter’s Vista AR project. Our Natural Language Processing algorithms enabled their team to skip a low value task (scrolling through reviews and collating the relevant information) and helped them make high value decisions (how do we respond to what customers think).
It’s about discovering what the problem is
A highly underrated part of data science is the problem discovery phase. People often imagine they can throw data at data scientists and get them to solve a problem with AI but in many ways this misses the point. If the data scientist went ahead with such a brief, in all probability the results they presented would be neither insightful nor actionable.
The issue is that before data scientists can generate satisfactory answers, they need to find out what the question is.
When organisations tell a data scientist about a problem, they’ll usually focus on the symptoms. The data scientist could go away and try and find the answer to those symptoms but halfway through the project they’d likely uncover a cause that costs more to solve. It would be a waste of everyone’s time and resources.
It’s rare that someone has privileged access to what a complex problem entails. So a data scientist can’t just listen to management’s view of the problem. Instead, they need to listen to a number of perspectives - often through intensive interviews - and synthesise what everyone is saying.
It’s about working through successive drafts
Finally, data science is not like software engineering. That’s a matter of breaking a problem down into smaller manageable problems, again and again, until you can quickly solve the individual parts and integrate them back up into a whole.
But data science has more in common with writing a novel. It doesn’t happen one perfect chapter at a time. It comes in successive drafts, each one more layered than the last, testing out variations of the same story until the characters feel real and the plot is seamless.
Data scientists work in a similar pattern. They need to test a number of ideas, evaluating them as cheaply and quickly as possible. And they need to iterate until they find the answer to one question at a time.
Organisations are on a journey here. You want to work in small incremental steps, starting with the most rudimentary. But gradually you’ll be able to go further and expand the number of cases you can solve with your algorithm.
The task of a great data scientist is to ask, ‘What’s the art of the possible here? And how can I take the first steps to help you?’ It’s a complex problem that requires a creative solution. And data science is the creative process that will get you there.
To explore how data science can help you create possibilities in your organisation, don’t hesitate to get in touch.