Built-in Graph Functions Accelerate Discovery Analytics

Posted on September 25, 2014 by

SteveWe often encounter analytics use cases from customers or prospects where the analyst wants to select a particular facet of the data and deeply analyze it. For instance, in the healthcare world, data includes patient (including genomic information), procedure, provider, billing and outcome data. When trying to discover new insights into the data, the analyst often doesn’t know exactly what she’s looking for, so she needs to answer several high-level analytic questions about the data to guide further exploration. (And, of course, today’s data is usually not just big — it’s Big Data, so performance and scalability are important.)

Continue reading

I work with superheroes!

Posted on August 15, 2014 by

MistiThe popularity of superheroes is timeless. And no wonder: It’s fun to think about a world populated by heroes with super-human intelligence, the ability to manipulate substances at the molecular level, and supersonic speed.

I’ll bet you can guess where I’m going with this. Maybe the summer movie season has put superheroes on my brain, but lately the things that Cray products are doing have reminded me of the superhero feats that excite kids (and grownups – you know who you are).

Some of this you’ve read in our blog posts over the last few months. There’s the “Beagle,” a Cray® XE6™ supercomputer at the University of Chicago, which is allowing researchers to test potential new drugs on the entire human genome, rather than only a portion of it – as they had previously been limited to. And they compressed 37 years of theoretical CPU time to 50.4 real-time hours. Kapow!

Continue reading

What Should Your Big Data Strategy Be?

Posted on August 7, 2014 by

AdnanIn previous articles, I discussed some common big data problems and causes of the problems.  In the final part of this series, we can now get to the icing on the cake – your big data strategy. Read on, my friends.

When do you have a big data problem?
Since our main interest here is in big data, a fitting question is when do you have a big data problem? The answer is not as straightforward as we’d all like but mostly because we need to have a paradigm shift in terms of how we think about the problem. This HBR article has some really good insight into how data visualization is helping companies understand complex consumer behaviors. The key is to think in aggregates and this is harder than it first appears because finer obvious details are lost. However, with enough data, more complex patterns begin to emerge. A good analogy is emergent theory, where patterns only become apparent and somewhat predictable due to the collective interactions of multiple different elements. On their own, each element exhibits very random behavior so you have to look at lots and lots of data together.

Continue reading