There is an analytics company that claims to have calculated “the best places to live”. West Perth in Australia takes the top spot in the whole wide world. The sceptic in me immediately thought, "whoever decided this probably loves picnics on the beach."
The company stressed that there was no single person behind the results. The ranking was not subjective - as would be the case had it been compiled by an editor – but objective as it had compiled freely-available unstructured and structured demographic, geographic and social media data. It then summarised it and applied machine learning and automated processes to decide what was and wasn’t relevant, all to “identify the characteristics that have the largest explanatory power to predict someone’s paradise.”
Those characteristics were restaurants and shopping, cost of living, safety and infrastructure, culture, family, education and career, and health. Then some local residents got into the mix and more machine learning was applied to well-known cities and the ranking was created.
It’s still a subjective process. Which careers were deemed important in this calculation?
That’s all well and good, but it’s still a subjective process. Which careers were deemed important in this calculation? Is Perth a good place for an aspiring aerial silks performer to be located? What is meant by culture? If I lived there, would I be able to meet new and established spray artists at outdoor art festivals? Does family mean child-friendly or does it mean that families with two mums or two dads would feel safe and welcome? Incidentally, Hebden Bridge, chosen by the “carefully selected” algorithm as the best place in the UK, looks like it would be a great place for two-mum families as it has been dubbed the lesbian capital of Britain.
How can you calculate atmosphere and vibe with a machine?
This is not meant to trash West Perth. I spent a day in the city and thought it was very lovely. But in the words of reggae artist Errol Dunkley; “Every man do his thing a little way different.” This is roughly echoed by Inma Martinez when she told me: “Your good may be my bad.”
Two people may be in the same city at the same time and have completely different experiences resulting in very different feelings about the place. How can you calculate atmosphere and vibe with a machine?
The company then encouraged me to use its ‘best place finder’ with my own set of criteria. Maybe if I set the parameters, it would be more in tune with what I like? However, I found it just too vague to be useful. In the restaurants and shopping category, there was no way to specify that I prefer independent shops to large malls, or that, in the nature category, I would take the beach over a lake any day of the week.
My ideal place was said to be Hood River in Oregon, USA, a town of 7,000 people located in an area known for apple production. Not quite. I would actually love to live in Santiago in Chile. One of my best friends lives there, but the machine doesn’t know that and there is no way for me to inform it of that fact.
This is just my feeling, but I think that likes and preferences are way too emotional and subjective to be algorithmically programmed.