Cloud Society: Facebook analytics is just the tip of the iceberg
Facebook introduced Graph Search last month - to much concern. But advanced analytics is an issue not about to go away
Since Facebook launched its Graph Search algorithm last month, pundits have been delighted, disappointed and horrified in equal measure. The "delighted" camp have been saying that, at last, the company has something with which to take on Google; more sceptical commentators have noted how the capability will not necessarily deliver on the hype.
Meanwhile however, articles have started to emerge which illustrate the potential for the tool – for good or ill. Such as Actual Facebook Graph Searches, which demonstrates how you can hunt for Tesco employees who like horses, or Italian Catholic mothers who like condoms. The possibilities are endless.
While may pundits (and no doubt their readers) have thrown their hands up in horror, the point that has been missed is, simply, it’s not just Facebook that can do this. Indeed, the whole premise of Big Data – to be able to use distributed processing and/or high-performance computing to run algorithmic analysis – is that it offers orders of magnitude greater insight than in the past.
You can hunt for Tesco employees who like horses, or Italian Catholic mothers who like condoms
It stands to reason. More data is publicly available than ever before, at the same time as organisations are collecting more and more personal data than ever before. A simple, yet compelling example was when I clicked a button saying I didn’t want to install a software product the other day. Had I not had a firewall running I would not have known that even the negative button-click sent a signal back to the vendor, which is right now logged in a database somewhere, indexed against my name, or my machine, or browser id.
They’re all at it – retailers, mobile providers, social networking sites, utilities, central and local government, indeed any organisation that has access to information is trying to get more of it, whether or not they know what to do with it yet. The insights that can be delivered may well be highly useful and beneficial.
Equally, they could be – as illustrated by the Facebook Graph Search, intrusive and potentially damaging. Consider – your mobile company has detailed records of the majority of your movements over the past few years. It knows exactly where you have been, and how fast you got there. There can be no better way of sourcing information about traffic jams than using the mobile network as a sophisticated set of sensors (shame this isn’t done). However, woe betide if you have been travelling significantly over the speed limit.
What’s to do? Even as the need for legislation is debated (and there are no clear answers here), organisations such as the Information Commissioner’s Office are relaxing the rules for ‘anonymised’ information sharing - in fairness, they have little choice. Meanwhile of course, the same organisation illustrated just how fallible it could be when it modified the cookie law implementation and left a hole big enough for a fleet of juggernauts to drive through.
While we can comforted that we live in a democracy which has the power to create laws to control any significant breaches of privacy or rights, it won’t be long before the data mountains around us can be mined on an industrial scale. The inevitable issues that will derive from this need to be understood and dealt with, and quickly.