Sometimes the real value of a conference is in the string of insights that come from it, not necessarily the presentations themselves. So it was at the Big Data event in Edinburgh (#scotdata).
For a start, Big Data needs renaming. Whilst we, in the communications industry, have often said that we deal with big data on a daily basis, it is nice to realise that everyone else is agreeing that the term big data is actually irrelevant.
Big data, according to more than one conference speaker, should be called, simply, data science. Why? Because, according to Suresh Pillai of ProSiebenSat.1, data science is about correlation. The real value of bi…, er, data science, is in the correlation of data sets. The more sets, the more value. The greater the variety, the greater value for different parts of your business. Finance needs one view into the information, marketing another and customer service another.
And while big data should be called data science, data scientists should actually be called data janitors, according to Steve Renals from the Alan Turing Institute. After all, they spend most of their time cleaning it and making it fit to use.
Pillai, who cut his teeth at eBay – whose knowledge of how to turn data into value is legend – re-enforced our theory that, when it comes to analytics, the boys in front are the retailers. The data science and correlation behind the Tesco loyalty card, for instance, is so sophisticated and successful that out of 60 million customers, only six received exactly the same combination of offers and coupons. Now that is personalisation.
Translate that into the potential value to be derived from, say, the 1.1 billion network impressions a day that Telefonica deals with (sites, calls, location check-ins) and the mind freezes for a second. The trouble, of course, is that 90 percent of data generated over the last 10 years has not even been captured. Let us hope that this will change.
Healthcare and the IoT, of course, were issues at the conference. The intelligent pill bottle that so impressed Finance Minister John Swinney at the IoT Scotland conference, was discussed again. This time, though, the context was slightly different but with the same goal. Emergency medicine – the midnight ambulance – is hugely expensive. Avoiding it makes lots of sense. One vendor has done an experiment in a care home (and a thousand sensors) to see if they could predict when someone was about to fall over. They did this by monitoring heart rate, temperature and lots of other things, and found they could. Being able to proactively do something to prevent a sick or elderly person falling over and triggering the midnight ambulance could save the health service millions. And keep people in their own home, which is cheaper, and better, for all concerned. The problem, according to Joe Duran of Fujitsu, is not digitising a process. That part is easy, it is the getting the approvals that takes the time and effort. Amen to that.
There were great insights from the conference and it was a pleasure to attend.
The only problem the organisers have now is what to call it next year. The Data Science and Correlation Conference 2016, perhaps.
Recent Comments