Puzzled by big data? Put it in context
In a recent study, Jeff Jonas, IBM Fellow and chief scientist for context computing (also recently named a Wizard of Big Data by National Geographic), explained how assembling a jigsaw puzzle inspires his context computing software portfolio.
The puzzle analogy presents a crisp and easy way to understand how context will drive better decisions and spot risk early. Think of a jigsaw puzzle with lots of pieces piled up on a table; this is much like data in our organizations, lumped into lots of distinct piles or silos. Analyzing data in isolation is like looking at a single puzzle piece, making it nearly impossible to make sense of the big picture. It’s only when we bring the puzzles pieces together that the emerging insight is revealed.
Essentially, this is how IBM context computing solutions work. They bring together different silos of observations and integrate them in real time to understand the larger picture without the need to train the system or use training data sets. Pivotal to this process is the ability to ingest and analyze any type of data from any source at high speeds with awareness of geospatial position. Order, timing and sequence of data points should not impact the ability to make sense of the big data or be necessary for the data to be of high quality. The IBM context computing engine makes sense of the noise and self corrects in real time as needed.
What does context computing look like in action?
- In a highly competitive industry with low margins, Moneygram uses context computing to determine if a customer may be a victim of a scam. They save hundreds of millions of customer dollars and have reduced consumer fraud complaints by 72 percent for better customer loyalty
- PEW Charitable Trusts implemented an election reform system using context computing. In a pilot of seven states, they identified 5.7 million eligible online voters to control the chaotic pace of polling stations on Election Day. Over 300,000 new voters were identified from the seven pilot states. Online voter registration saves 95 percent of the total cost of voter registration
- The Rochester police department improved efficiency by 95 percent and reduced the investigative process from days to minutes
- An insurance company constrained by the manual and ad hoc approach to detecting and investigating potential fraud verified, in just weeks using context computing, two criminal fraud rings that took years to identify using manual methods and also located a previously unknown suspect fraud ring
Of course IBM builds privacy into the context computing platform. Today, with the cost of data breaches up 14 percent, security and privacy controls must be built into big data technologies. As we derive context from data, it means sharing information and source of data across different functions internally and externally. Effective information sharing can only happen when individual identities, personally identifiable information and private data is kept safe. The IBM context computingplatforms protect and facilitate information sharing without breaking the context.
IBM context computing turns big data on its head and introduces a new physics where more data means better observations, faster. Think back to the puzzle example: as you fill in more pieces, the ability to sense, orient and respond gets faster and more accurate. The goal is to spend more time pursuing real opportunities and less time chasing false positives and negatives.
With that said, what does your puzzle look like? Tell us what solutions context computing could provide for your business in the comments below.
For more reading:
- Jeff Jonas’ blog
- IBM context computing
- IBM Stream Computing
- Forrester Wave- Big Data Streaming Analytics Platforms