Anyone who hated mathematics and statistics at school has probably turned something of a deaf ear on all this talk of data analytics. The second breath in any conversation centering on ‘big data’, today we know that data analytics is the term the IT industry uses to talk about enterprise-level machine based intelligence. But unless you harbor a proclivity for statistics and probabilities studies, how could this dry and dusty subject possibly be interesting? It either has to sound sexy and exciting, or it has to start to impact your own business or your personal wellbeing – or both.
Donald Yeoman is manager of NASA’s Near Earth Object Program Office. Yeoman has said that it is likely that an asteroid as large as 50-meters in circumference is likely to hit the Earth somewhere around once every 500 to 1000 years. Yeoman’s studies have been cross-referenced in light of the work being carried out by IBM’s data analytics division. Do you see where we’re going with this?
An informational hors d'oeuvre
“We have been using geospatial data analytics to look at how asteroids interact with other asteroids when they hit each other,” said Jeff Jonas, an IBM Fellow and the firm’s chief scientist of context computing – a man who also spends time on the board of the US Geospatial Intelligence Foundation (USGIF). Jonas uses this ‘informational hors d'oeuvre’ to whet audience appetites before delivering more mundane sounding applications of big data analysis.
Back down here on Earth, we are all discussing how data analytics is about to impact our household items when they all eventually connect to the Internet inside the so-called Internet of Things. For ‘household’, also read wearable devices and automobiles. There is arguably good reason to start connecting our cars’ sensors to our mobile devices so that we know when tire pressure, water, oil and other ‘vital signs’ need servicing. Logically onward from cars, we move to aircraft.
Speaking in conjunction with his firm’s work carried out with IBM’s analytics division this month was Larry Volz, who is chief technology officer at aerospace manufacturing company Pratt & Whitney.
Old (prescriptive) data is like looking in a rear view mirror
“Pratt and Whitney had been looking at what happened to the motors inside our engines i.e. how they performed and where they went wrong. But it was always like looking in the rear view mirror. We wanted to be able to use big data analytics to move from a prescriptive model to a predictive one. But to do this any firm (including ours) needs to reengineer entire business processes and certifying new products under these new models in the aerospace industry is not simple -- so there is still a lot of work to do,” said Volz.
But big data analytics will not just be put to work inside the aircraft’s avionics, it can work in the cabin too. Imagine if inflight cabin crew could walk through aircraft with a personalized travel schedule for each customer and, when a delay occurs on a connecting flight, be able to print off new boarding passes and hand personalized terminal direction advice to each customer without having to even question them first about their plans. This kind of software application development is not far away and we might even call this the ‘consumerization of business’ if you will.
Insight comes from big data analytics, but we also need a channel to visualize our results - Photo... [+]
IBM has talked about new scenarios in line with the release of a new cadre of analytics products which include: IBM DataWorks – a set of cloud-based data refinery services to help users shape, cleanse, match and secure data; IBM dashDB – a cloud-based data warehousing and analytics service with in-memory technology; and IBM Cloudant Local - an on-premise edition of the cloud database-as-a-service that enables a fluid hybrid cloud data layer that spans private datacenters, mobile devices and third-party cloud providers.
Analytics pays back $13.01 for every dollar spent – 1.2 times more than it did three years ago, according to Nucleus Research, September 2014. Yet says IBM, further studies show that data cleaning and preparation takes approximately 80 percent of the total data engineering effort. The firm cites these figures as some justification for the functionality it has developed in its most recent set of product launches.
IBM has also announced new high-speed analysis software which will typically be used for criminal investigations. Designed to uncover hidden criminal threats buried deep inside massive volumes of data, IBM i2 Enterprise Insight Analysis (EIA) claims to be able to find ‘non-obvious relationships’ masked within hundreds of terabytes of data and trillions of objects in just seconds
Six degrees of separation
According to IBM, “EIA analyzes huge amounts of disparate data to discover weak-signal relationships that reveal the true nature and source of an attack. The solution unravels these hidden connections that can be divided by as many as six degrees of separation between disparate sources – from corporate records and social media chatter to data accessed by remote sensors and third-party applications. As developments unfold, EIA provides always-on recommendations that proactively alert analysts to new related abnormalities at the speed of attack.”
Although this example sits in the field of fraud and criminal detection, this idea of six degrees of separation in data analytics brings us right back to asteroids. Say a user enters his or her name and age into an analytical database, the analysis engine can start to tell you what year that person was born in using related logic information. As the user starts to offer up a little more contextual information about themselves, the machine learning element of the data exploration engine fires up and starts to look make predictions based upon some level of certainty, which is better than simply guessing.
Looking at which way colliding asteroids will behave is all about knowing their size, mass, trajectory and speed – but this is space and almost all of these figures will be estimates based upon measurement and calculations, but also on intelligent reasoning. Do you know the magnetic force architecture of an asteroid? No, nor do we and nor does NASA and nor does IBM, but contextual information analysis can start to guide us towards the best and most likely answer at any point in time.
Everything goes better with bacon, and data
Does all this extra prevalence of data and information analysis make our lives better? The answer is might be yes if it makes our planes fly safer and our cars operate more efficiently, but it will come at a cost and someone is selling this software. As suggested on Forbes yesterday, all this data analytics and information insight will ‘probably’ leave us all better off. Or as VP for business analytics at IBM Eric Sall put it, “Everything goes better with bacon -- and everything goes better with some data too.”
