The recent Gartner BI Summit in Los Angeles provided an opportunity for Enterprise Apps Today to meet with a wide range of business intelligence vendors to find out what’s new on the business intelligence front. Here are six of the more interesting products seen at the summit.
QuartetFS Analytics Cube
QuartetFS has developed what it calls an in-memory OLAP cube which is aimed at solving Big Data problems. The basic idea is to eliminate the time wasted waiting on query results to materialize. Instead, QuartetFS updates its analytics cube whenever new facts come in. Thus it is constantly recalculating key factors against the flow from marketing and other data streams.
QuartetFS Managing Director Allen Whipple gave the example of a banking challenge centered around what the financial world terms value at risk. This is essentially how much capital to set aside to lower risk. If they get it wrong, the Fed can bypass them and that comes with a hefty penalty; the Fed can double the amount a financial institution needs to set aside -- which can endanger profitability and share value.
“Most finance companies review their value at risk once or twice a day as it takes several hours to run the query,” said Whipple. “Our customers are able to continuously predict it based on aggregated constructs. This enables them to be far more certain of their numbers.”
According to Whipple, the product is good for logistics and cost-of-business decisions in e-commerce.
Logi Info Active Analytics
LogiXML brought along a customer, Kevin Dodson, from outsourcing account/payroll firm Baker Tilly Revas to showcase a new addition to its product lineup called Logi Info Active Analytics. Baker Tilly Revas created a client portal from LogiXML that enables it to provide its clients with business intelligence capabilities such as reporting and dashboards. They can then drill into the data and perform their own analytics. This client portal ties LogiXML into six core apps that the company offers its customers.
“Approvals and rejections can be done operationally, rather than having to wait for a management decision,” said Dodson. “It isn’t the usual BI product in that it doesn’t need a data scientist to understand it. Regular business managers pick it up immediately.”
Revolution’s Business Intelligence with Statistics Edge
Jeff Erhardt, COO of Revolution Analytics, positions his company’s product as Red Hat for statistics. His company is the commercial backer of the open source statistical language known as R. Erhardt believes R is taking over the fields of research and statistics. He points out that companies like Facebook, LinkedIn and banks (for trading models) already utilize it to derive knowledge from Big Data.
Erhardt is also quick to point out that Revolution is not a business intelligence provider.
“We make it possible to provide sophisticated stats and analytics to BI,” he said. “We don’t just help people look back at what has happened but use mathematical modeling to forecast the future.”
eBay, for instance, uses R in its real-time recommendation engines which pick out suitable recommendations of additional purchases for customers out of its massive database of transactions. Similarly, Facebook uses it to track possible new connections by analyzing and comparing the friends of friends to find matches. Another example: a hedge fund importing Twitter feed sentiments as an additional element for its trading model.
SAS Visual Analytics
SAS just introduced SAS Visual Analytics, which provides in-memory analytics on commodity blade hardware. The idea is to make it cost effective to scale by adding a couple more blades. It is basically a combination of SAS analytics running in-memory along with Hadoop support, visualization and mobile functionality.
The SAS LASR Analytic Server supplies the in-memory architecture and local storage is provided using Hadoop.
“While there are lots of visualization and in-memory products out there, this tool combines them all to visualize Big Data,” said Jennifer Marchi, product marketing manager, SAS. “It can provide visualizations that work even on billions of rows of data.”
One feature is some intelligence to select the right means of visualization. With lots of data points, a scatter plot can become no more than a giant blob. Therefore, the software might suggest a heat map as a better alternative.
For instance, a large department store like Macy’s has billions of individual items within its inventory. That doesn’t show up well in a bar chart. Therefore, SAS might display instead a ranking of the top 10 or top 100 items ordered against such criteria as ascending or descending profits.
HiQube’s Simulation Play
HiQube is another company taking the approach of tacking functionality onto business analytics. While Revolution focuses on the statistics angle, HiQube’s core competency is math-based simulation. It tackles problems like, “If these are our objectives, how do we get there?” To accomplish that, it utilizes what-if analyses and simulations.
Case in point: a utility that installed tens of thousands of smart meters at customer sites across its network. Instead of the meter being read manually once a month, the utility now has data coming in continually from every meter. This enables the utility to fine tune its pricing structure from a simple off- peak or on-peak pattern to having many different rates across the day, evening and weekend. By informing customers of this, they can decide not to do laundry at 5 p.m. but instead wait until after 8 p.m. when rates are lower.
“The goal is to normalize usage across the day versus building new power generation facilities,” said Michael Kidder, senior vice president of HiQube.
Pentaho Finding Patterns in Big Data
Pentaho is seeing plenty of usage by companies that already have a business intelligence platform but chose to embed Pentaho into it to find patterns among Big Data sources, said Rosanne Saccone, CMO of Pentaho.
The company chooses an open-source approach to business intelligence, making use of Hadoop and Java. The latest version of its software (4.5) is focused on visualization and user friendliness, while adding Big Data and data discovery capabilities.
Saccone made a good point about integration. Many companies suffer from data mart sprawl. They develop different marts for each business unit. Consequently, each unit may create visualizations that offer a completely different picture of what is happening with the latest marketing campaign. The CFO, visualizing his own data mart, sees things one way. The CMO’s mart highlights a different situation. This is the problem with having separate tools sitting on each data mart.
“Pentaho gets data from all the marts and unifies it to provide a better version of the truth,” Saccone said. “A platform approach is better than a power tool approach.”
Drew Robb is a freelance writer specializing in technology and engineering. Currently living in California, he is originally from Scotland, where he received a degree in geology and geography from the University of Strathclyde. He is the author of Server Disk Management in a Windows Environment (CRC Press).