Sergey Nivens - Fotolia

IBM adds cloud analytics platform for utilities; de-anonymizing data

This news recap looks at IBM's announcement of a cloud analytics suite for energy and utility companies and a study showing that anonymized data sets can easily be reidentified.

IBM last week announced a new cloud analytics platform with a set of tools aimed at energy and utility companies, offering functionality designed to enable them to track and analyze things like energy production and distribution in an effort to improve business decision making. Also in this news recap: The anonymized data sets used for research purposes aren't so anonymous after all, according to a study by a team of researchers at MIT.

IBM powers up cloud analytics for energy industry

IBM's new Insights Foundation for Energy product offering is delivered on its SoftLayer cloud platform and packages together data integration, advanced analytics and data visualization tools. IBM said in a press release that the cloud-based platform is intended to lower the financial barriers that are holding back some energy producers and utilities from becoming more data-driven.

As part of the package, users get a suite of built-in analytics tools, including ones to help them manage predictive maintenance projects and assess system-wide risks in real time, IBM said.

Interest is growing in smart grid technology and new analytics capabilities that can aid utilities in better managing their power generation, transmission and distribution assets. Proponents say that by analyzing how energy is produced and delivered, and using data to improve equipment maintenance, utility companies can improve efficiency and reduce costs.

The IBM rollout comes at the leading edge of what could be a growing cloud analytics trend. Last year, several major analytics vendors announced new cloud-based products, and analysts are increasingly saying that the cloud is becoming a mature option for delivering analytics.

Study: Anonymized data sets aren't so anonymous

The cornerstone of much analytical research, especially medical and scientific studies, is the anonymized data set. And as corporations have begun basing more of their operational decisions, like marketing campaigns, on data, trading anonymized user data sets has become a big business.

But these data sets aren't as anonymous as promised. A new study published in the journal Science, conducted by a team of MIT researchers that included Alex Pentland, shows that it often is a simple matter to restore identifiers to previously deidentified data sets.

First, the team of researchers obtained an anonymized set of credit card data. They said that they then were able to identify unique buying patterns, even when usernames and account numbers were absent. From that information, they were able to tie purchases back to specific individuals.

The study was published a week after a user of the website Reddit said he had used an anonymized data set of information on New York cabbies to identify drivers who might be devout Muslims, based on their driving activity around that religion's daily prayer times.

The authors of the Science study wrote that their findings show a need to rethink how potentially sensitive private data is handled.

This week's dataviz

There has been a lot of talk about measles in the past few weeks, with a significant outbreak at Disneyland in California and a handful of prominent politicians weighing in on whether vaccinations against the disease should be mandated. This prompted the World Health Organization to look at vaccination rates around the world. Check out the WHO's data visualization to see where different nations stood on measles vaccination rates between 1980 and 2013.

Ed Burns is site editor of SearchBusinessAnalytics. Email him at [email protected] and follow him on Twitter: @EdBurnsTT.

Next Steps

How to be happy with BI and analytics in the cloud

SaaS model solves some big analytics headaches

What kind of tools make up Oracle's cloud analytics platform?


Dig Deeper on Advanced analytics software