To some, the Spike Jonze film Her is a ridiculous watch. The idea that a man could become infatuated with a machine is surely a thing of science fiction. To others, it’s an almost dystopian look at humans’ relationship with technology pushed to the brink.
Whatever your interpretation, one consistent theme is that the relationship between humans and data is increasingly complex. A lot of data generated today is about humans but to what extent does the analysis of big data help and inform about the human race, and how can the worlds of big data and humans be best intertwined?
Today, data is being used to improve operations, help with decision support and invent services that stay ahead of changing times and customer needs.
One might argue that the Holy Grail of data analytics lies in understanding and modeling human behaviour to accurately predict how it will change when variables and circumstances are altered – ultimately allowing it to preempt human behaviour.
Is this achievable? As people share more and more of themselves online, the increasing quantity of data is being used by systems to get better at analyses. But we’ll only be able to have greater influence on human decisions when all relevant data is integrated and analysed to the right level of granularity. Even then, knowing when or how to intervene to encourage a particular outcome will remain difficult.
Still, big data research persists in trying to represent and influence human behaviour using data. Why? It’s the nature of any research: to find out what’s possible.
Xerox has investigated a number of areas in this field, including crowdsourcing, smart parking and sustainable commuting.
Understanding how crowdsourcing markets function (such as Amazon Mechanical Turk) and the needs and behaviour of workers within them to a useful degree is currently next to impossible. But analysing the data workers generate can allow them to search and make informed choices on the tasks they choose to do, and understand their work better.
By changing the cost of parking rates based on live traffic data (smart parking), people are encouraged to park in a way that reduces congestion. Although price changes are driven by data analytics, the reason it’s successful is because the information is available to drivers, who can learn about street parking prices from the web or smartphone apps, and make informed decisions about where they park versus the cost.
Gathering data from toll roads, pedometers and ticket machines to study how people travel is a means to suggest better transport alternatives, like getting people to walk and cycle more and reduce the use of cars (sustainable commuting).
Put to good use, data analytics will enable decision-making and ease some everyday frustrations. The opportunity therefore is there to sign over more of these menial, yet human, decisions for technology to either control or advise people and help them make a choice.
However, while big data can help influence choices, it doesn’t provide a complete understanding of why people actually make them.
Lost in translation
Think of the amount of data that social media gathers on people’s lives everyday, yet even something simple such as Facebook likes are not all the same. Some are made almost by reflex, others reflect a deeper appreciation of a post, and others may be less about the content and more about the relationship we have with the person.
In a similar vein, it should be obvious that not all messages are the same; a note conveying some valuable information or sentiment from someone important to you is far more impactful than the lunch choice of an old colleague.
Automated social network analysis, while interesting and even rather sophisticated, produces a perspective that doesn’t accurately reflect reality because it cannot capture the rich interconnected human experiences that span on and offline.
The consequences of this may be irrelevant in many cases, moderately annoying in others, or have significant negative impact on some individuals.
This is not to say that these technologies will not develop and improve. But data analytics will ultimately stop short of being able to read, reflect and understand to the sensitivity level of humans.
Customer care is an example that demonstrates this shortcoming. Businesses would love to get customers to self-serve more – and we can already see this growing significantly through internet banking, commerce, and online troubleshooting.
But businesses have to pay a lot of attention to how electronic interactions work, and some of this comes down to giving interfaces ‘soft skills’ – being pleasant to use, and using the right sort of phrases.
Having an interface that works effectively at answering queries is a very reasonable ambition, but creating one that could convincingly read emotions and converse with humans will not exist in our lifetime.
Keeping the human touch
As the study of big data continues, computers will become the masters of tasks that involve calculation, or that can be reduced to a series of calculations and operations.
But for all the incredible advances of data analytics, the best interpreters of human behaviour are, and always will be, humans, because social interaction and understanding often cannot be reduced to computable data or information.
Ever shouted at technology before? People respond with anger and frustration when they come up against devices that fail to understand them. Certainly in Her, this quest for understanding is explored in great detail – without resolution.
>See also: Big data and mapping – a potent combination
Of course, people also get frustrated with other people. But the difference between technologies and people is the potential for understanding. If people care, they can often find a way to reach mutual understanding.
Unique discoveries are made every day in big data analytics and these should be put to good use for all. But let’s not forget that humans and computers have different capacities and abilities, and value how they complement one another.
Human capacities for empathy, aesthetic appreciation and reflection are some of people’s greatest strengths, but are notoriously problematic to compute.
Big data analytics can be a wonderful tool, but as tools, they are designed to serve people. Humans decide when it’s appropriate and ethical to use them.
Sourced from Dave Martin, Xerox