Next generation of big data analytics emerging

The next step for business intelligence clearly involves an increase in the amount of information being examined. This much has become clear from the discussion around big data, with algorithms that accommodate large and varied sets of content becoming popular across industry lines. This is not a promise of effectiveness, however. Simply jumping aboard a trend and buying software will not instantly make a company into an analytics powerhouse. It's up to leaders to figure out what data means to them and make the change. It won't happen overnight, but the change may be just as impressive as advertised.

Collecting the right type of data
ReD Associates' Mikkel Krenchel and Christian Madsbjerg, writing for Wired, recently gave an explanation of why some programs may fail, even when they have been equipped with all the power afforded by big data. The authors suggested that for all the power of algorithms, it helps to equip these programs with facts drawn from life, "thick" data as opposed to raw information. This means calibrating queries based on the observable way in which people act instead of drawing a projection of the whole situation from a stream of numbers. It entails combining logic with the powerful streams of information that are now available.

The authors gave Google Flu Trends as an example of their theory. They noted that this system, designed to pinpoint flu outbreaks through search terms, misclassified words that had to do with winter, erroneously placing them in a cause-and-effect relationship with flu symptoms and predicting many more cases than actually took place. The example shows that a little know-how from a guiding human hand can create a better use for an algorithm than would be possible if the computer was simply left to run alone. Discoveries such as this are powerful proof that business intelligence and big data algorithms are here to help analysts, not replace them.

New horizons for data use
While the process of collecting and using data may fail without oversight, plenty of companies are ready to make it part of their operations. The next few years may prove to be one big ramp-up. TechTarget recently reported that one large medical firm is growing its processes slowly, with a gradual addition of new capabilities built in. The plan will take five years to be fully operational, but it has already begun, with data flooding in. The presence of such a refined strategy also means the products in question were picked carefully and intelligently.

Submit to our blog