Hadoop Enables Business Analytics

Okay, we’ve got the data warehouse infrastructure, distributed file systems, a database and plentiful CPU cycles. So what’s most needed by those still unable to find big value from big data? Analytics!

While some well-resourced organization can throw technical talent at uncovering the value in their Big Data, other users are struggling to find analytic technology that takes full advantage of the richness and scale of the Hadoop ecosystem. Several examples of what is required for problem solving on such a massive scale are:

• Data Visualization of massive datasets. A billion points in a scatterplot isn’t very useful if you don’t have a billion pixels on your display. (hint: you probably don’t)

• Supervised variable selection over a decades’ worth of data with 30-60,000 attributes. For example, financial service organizations have typically been constrained to 3-5years and fewer than 5,000 attributes.

• Mobile operators are capable of generating massive datasets from the switch-level diagnostic logging capabilities of modern 4g wireless infrastructure. What can we learn from this dataset looking for early warning signals of conditions that lead to unhappy customers?

• Insurers are looking at telematics datasets generated when customers opt-in for measurement with the promise of driver-specific insurance rates correlated to safe (or unsafe) driving patterns as shown in the sensor data.

In this session Paul Kent, Vice President of Big Data at SAS, will show examples of how members of the Hadoop community can provide their enterprises with critical insight from Big Data assets.