Intelligent Enterprise

Better Insight for Business Decisions

RSS
Webcasts
Whitepapers
Subscribe
Home

August 10, 2003

Analytics in Manufacturing: The Collaboration Enabler

The quest for profitability has businesses in manufacturing industries on the hunt for analytics that serve collaborative execution models. Will BI providers step up?Or will solutions come first from ERP?

by Naeem Hashmi

Continued from Page 1

Unstructured Data

A significant portion of product design and development involves unstructured data, such as computer-aided design diagrams, discussion documents from meetings and conferences, and so on. And so, we have another challenge in front of traditional BI tools: Can they analyze unstructured data and perform content mining? Can they build taxonomies and use them to analyze a knowledge base to derive new rules and thereby enable users to actively participate in driving the entire product life cycle?

Good document management systems exist in the market, but these manage only unstructured data. Integrating this data with other resources that traditionally feed business analytics is a complex proposition. ERP providers such as SAP, Oracle, and PeopleSoft offer product life cycle management (PLM) solutions that use portals to integrate underlying business and product applications with content management and analytic services. The result is technology that can help organizations improve product quality, reduce planning time, and bring products to market faster.

Data Mining Processes

To meet the challenges posed by trends in manufacturing, data mining will require some reorientation. Today, most data mining activities are still left to statisticians, who model, train the data, and crank algorithms through powerful computers for hours and days to discover anomalies hidden in large data sets. Then, the findings are presented to business decision makers, who identify what actions the organization must take. This modus operandi works fine and will certainly continue; however, it is a time-consuming process — and time is everything in today's business environment. Decision-making must happen as quickly as possible.

Three important trends are emerging in the data-mining field, primarily to address decision-making speed:

Online analytic mining (OLAM)
Databases embedded with data mining engines
Integrated business data mining engines from ERP providers.

All three trends bode well for BI service providers. We can look a little closer at each.

The OLAM connection. OLAM is quite appealing to manufacturers with tons of data waiting to be analyzed. (See Figure 1.) With powerful end-user workstations becoming less expensive, we are seeing products emerge such as PolyVista, which offers an integrated OLAM system. PolyVista provides sophisticated algorithms that automatically mine multidimensional data and then present hidden anomalies in an interactive, intuitive fashion to help decision makers understand the product's findings quickly and delve further. The beauty of PolyVista is that it enables business users to make quick, intelligent decisions without having to learn statistics, understand complex data extraction methods, or go through training exercises. In other words, PolyVista and other new tools are bringing data mining and knowledge discovery from the back office to the front office.