Curiosity and exploration in Data
If you’ve ever picked up a book about tech, chances are the author’s story starts something like: “When I was [7] years old, I learnt to programme on my father’s [Commodore 64]…”. These stories are meant to share the moments that sparked their passion for the technical field or concept.
(They aren’t meant to make you feel like you’re decades behind in your tech/programming journey, but they might.)
(They aren’t meant to make you lament the absence of disposable income in your family… But they might.)
At the core of their childhood story, however, is the most potent of human attributes: Curiosity.
Curiosity is defined as a need, thirst, or desire for knowledge. The concept of curiosity is central to motivation and a prerequisite for exploration and innovation. Decades of psychology research on curiosity has shown that it can improve intelligence and fuel divergent thinking and build the stamina and adaptability needed to propel us to acquire the resources needed to transform ideas into action. The latest studies have prompted prestigious journals to tout the benefits of curiosity and businesses to specifically hire for it.
If you work in tech or Data, you might recognise the little sparks of inspiration during your latest Youtube spiral, choosing the perfect habit tracker on New Year’s Day or that Machine Learning course that blew your mind. If you can’t think of any, just hang around long enough on Beyonce’s internet, and you’re bound to come across impactful visualisations, impressive dashboards, funky AI applications that would leave any aspiring Data Person in awe of what is possible. No matter the subject, this heady mix of novelty, complexity, surprise or even incongruity has all of us techies hooked.
All this frenzied arousal can have a powerful impact on us and our careers. For the Data enthusiast, it might just drive us to spend our Learning & Development days brushing up on our dashboarding skills or starting a nerdy Pinterest account with our favourite dataviz inspirations. An AI fan on the other hand, might channel their energy towards learning the latest in Python modules, reading academic papers for algorithmic inspiration, or searching the web for the latest AI software. This ongoing quest to learn, to be useful, to be better powers our growth and our maturation in our careers.
For a Black person in Data, following our curiosity can come with its own surprises. Perhaps you’re passionate about advocacy and decide to direct your data skills towards the latest equality & diversity initiatives at work. One look at the dataset from HR, however, and you’re reminded that you are one of a handful of non-white entries. Your data experience tells you it’s not enough to do statistics upon. Your life experience tells you it’s enough to be a statistic.
Perhaps you’re more interested in applying yourself to a technology project – the cool stuff of science fiction books. MIT Media Lab researcher, Joy Buolamwini did just that. She set out to build a mirror to overlay inspirational figures over her own face and found that the facial-recognition software she planned to use misidentified and misgendered darker-skinned females.
Joy’s Gender Shades project highlighted a persistent problem in tech and Data. Those creating the facial-recognition systems were probably just as excited as I would be to be working on an innovative piece of software. However, such blinkered enthusiasm can be dangerous – allowing exclusionary and discriminatory practices to be amplified and propagated through code.
To build good Data products, we need to channel our energy, motivation, and curiosity outwards too – to the people our code and analyses serve. Rather than spend weeks working on a model or analysis in isolation, spend time understanding the perspectives and motivations of your users and stakeholders. Is the data you plan to use appropriate and representative? Do you have rigorous methods in place to test your models? Will your outputs be usable? If so how and by whom?
Even the most advanced and accurate machine learning model is useless if it doesn’t solve their specific problem. By asking questions early and often, you’re more likely to avoid common pitfalls in data analysis, identify creative solutions to their problems and produce outputs that they will value.
No matter your career or Data skills, paying attention to your internal map of inquisitiveness will guide you to the variables that matter to you and others, fostering healthy relationships and more satisfying work.
So, tell me, what are you curious about today?