LAS VEGAS -- How best to deliver big data analytics to users was under discussion here at this week's TDWI BI Executive Summit. Presenters and attendees alike looked to chart the way to successful analytics initiatives that connect big data and business value. Such initiatives would work on the mountains of unstructured data gathered from new Web-based applications, cloud computing and far-flung sensor networks.
The key question business intelligence (BI) professionals must ask hasn't changed with the advent of new approaches, said Barbara Wixom, associate professor at the University of Virginia's McIntire School of Commerce. The goal, she said, is to answer the question, "How do we get the data to the users?"
"There is no value created without use," said Wixom, who shuns the term "big data," saying it lacks definition. She does agree, however, that "data is changing." These changes require data professionals to redouble efforts to find business value in data architecture.
"The quantity of sources themselves is becoming overwhelming. We have more data sources popping up every day," said TDWI conference keynoter Evan Levy, vice president of business consulting at SAS Institute Inc. in Cary, N.C. The key problem may be figuring out how to move that data around the organization and to deliver it to business users, he said. At the same time, data architectures have become a bit of a black box. "I have a rat's nest of code going back and forth. What do I
Levy said data professionals will be best served by taking incremental approaches to dealing with the influx of data. He also told attendees to take cues from established manufacturing supply chain applications, which deliver final, uniform products working from diverse raw materials.
Sabre on Hadoop and BI
The growing supply of Web data challenges conventional data warehouse and analytics delivery planning, according to Jessica Thorud, director for enterprise travel data warehouse and business intelligence solutions at Sabre Holdings. She said the travel giant is moving to alternative data strategies due in part to the huge volume of shopping data it is gathering. At Sabre, big data and other technology implementations are accompanied by thorough business planning, she added.
"We know it is coming, and we know how to do it," said Thorud, who is continually looking for new and better ways to connect modern Hadoop big data tools to enterprise data warehouses and analytics systems. "We hope the [Hadoop] tools continue to evolve so we can connect to the BI tools," she said.
While it's early in the big data era, Thorud predicts that her company will one day deliver big data analytics to customers, especially those focused on marketing applications. "They want the insight. They have ideas on how to use it in decision support," she said. To deploy Hadoop, Sabre selected the jointly produced Oracle-Cloudera Hadoop-based Big Data Appliance.
Thorud's comments came as she took part in a summit session on delivering innovation in travel through data and BI products. She encouraged attendees to focus on usability and design when delivering BI products. She also reminded them to "know your customers' needs and capabilities."
Which data initiative is first?
Like Thorud, academic Wixom noted that marketing-focused applications have garnered significant attention among new and innovative analytics initiatives. She told the TDWI audience that she had studied the early efforts of data warehousing, finding that back then, marketing also led the way in adoption.
"There is lots of buzz around big data. Different people have different concepts," said conference attendee Masood Ali, information management architect at the Royal Bank of Canada, who came to the TWDI summit in part to help determine the correct big data strategy for his organization. Despite the hype, he sees greater use on the way. "Big data will soon become normal data," he said.
"The key is in finding the value. The important thing is to build for purpose, so big data has value," Ali said.