What is Data Mining?
The term Data Mining has been used to describe everything
from artificial intelligence, to the CIA or FBI connecting the dots on terrorist
attacks, to a person manually pouring through tables of arcane data.
More properly, in the context of PCA's Data Mining Consultant Services,
data mining is the process of analyzing data to find hidden patterns using technologies
such as OLAP, Cubes, Pivot Tables, and advanced algorithms such as Naïve Baysian,
Microsoft Clustering, and Neural Networks, and applying the results to actionable
business initiatives. Data Mining
is also referred to as Knowledge Discovery in Databases (KDD), Smart Databases,
Intelligent databases, or Predictive Analytics
SQL Server relational databases are at the root of nearly all data mining efforts that PCA gets involved, though this
is not a requirement. Much of the underlying data we use to data mine comes from custom and off-the-shelf business software such as
Customer Relationship Management (CRM) or Enterprise Resource Planning (ERP/MRP) applications, along with Insurance
Software, Financial Software, and Web Log Data. Our Data Mining Consultants primarily use tools from the Microsoft
Business Intelligence platform delivered as part of SQL Server to deliver data mining solutions.
SQL Server Data Mining is built into the SQL Server Business Intelligence (BI) platform, fully integrated
with Analysis Services, OLAP, Integration Services and Reporting Services. SQL Server Data Mining has become the most
widely deployed data mining server in industry today, making it an appealing platform for data mining consultants to
tackle small, medium and enterprise level data mining needs.
Tenets of Effective Data Mining Solutions
- Good Data. As the number of stories you can put
on a building is a function of the strength of its foundation, the capabilities
that a company can derive from data mining is limited to the quality of the underlying data model. Effective data mining
solutions must be built on a properly structured, representative set of business data.
- Purpose Built. Customer data mining is the application
of descriptive and predictive analytics (such as clustering, segmentation, estimation, prediction and affinity
analysis) to support the marketing, sales or service functions such as customer segmentation, cross-sell or
- Ease of Use. a user interface suitable for company
users (such as campaign, segment, product, sales or service managers) to perform analyses is critical to an effective data mining strategy.
- Intelligent. robust data mining algorithms able
to provide reliable and scalable insights into a variety of types and volumes of customer data.
data mining produces greater results when the ability to access data from heterogeneous sources, particularly those with information about customer
interactions and transactions (such as customer data warehouses, call centers, e-commerce or
Web-site-tracking systems), as well as third-party data providers that supply customer-related
information (such as demographic or market spending information) is an option.
- Accessible. ability to make the results of the
data mining analysis available to the point-of-need i.e. senior executives, line-of-business managers and salespeople.
Primary Users of Data Mining Solutions
Data Mining is used most commonly by executives who are involved in strategic and tactical decision making as well
as line-of-business managers responsible for cost reduction initiatives. Marketing Managers use data mining to
gain insights on competitive intelligence, to explore untapped customer and market opportunities, to support product
and service positioning and pricing decisions, and to perform detailed marketing campaign analysis. Sales and
Customer Service Managers who responsible for tactical decision making use data mining to support sales forecasting,
direct marketing, and customer acquisition, retention and extension (cross/up-sell) purposes. Operational managers
use data mining to identify inefficiencies and bottlenecks in complex supply chain, production and distribution networks.