PolyVista® is business intelligence software that
extends the quality and value of business analysis and
reporting. While Reporting systems generally deliver analysis
in only one or two dimensions at a time, PolyVista in
contrast, automates analysis across any number of dimensions…
simultaneously! Thus PolyVista addresses a broader
range of business requirements, from basic reporting queries
to understanding complex relationships involving many
attributes and large data volumes that include structured/numeric
and unstructured/text data. PolyVista has been designed
for the business user rather than statistician, to enable
businesses to fully leverage their data and identify important
new issues or opportunities that too often remain unknown
and unexploited.
Discover
the Value
The PolyVista solution is unique in its ability to assist
the user in discovering "insight gems" hidden in the data.
In our model, new opportunities and potential issues are
explicitly brought to the users' attention, thereby encouraging
the validation and learning process. PolyVista's discovery
techniques actually guide the users in their quest for
new opportunities or process improvements. It is through
this process of discovery that companies will truly be
able to re-invent themselves and ensure their success
into the future.
By seamlessly integrating the rich functionality of On-Line
Analytical Processing (OLAP) directly with interactive
2D/3D graphics and the automation of data mining, PolyVista
provides an exciting discovery focused experience for
the end user. Our "Discovery" (data mining) algorithms
are designed for simple configuration and are tuned to
leverage the unparalleled performance of OLAP databases.
These algorithms can surface business insights that represent
discovered anomalies, relationships, or business rules
hidden deep within the OLAP database. These insights may
take the form of a better understanding of your customers
and of their buying patterns. They can also identify new
business opportunities or may reveal anomalies and inefficiencies
in your business processes. The discovered "insights"
are presented to the user for business value assessment
and can, in themselves, represent important, actionable
business knowledge, or form the basis for further analysis
and verification - all seamlessly supported by our tools.
The Whole is Greater than the
Sum of the Parts
PolyVista uniquely and seamlessly combines four powerful
technologies to assist business users extract value from
corporate data repositories.
 |
|
An
OLAP engine (industry standard) |
|
|
Text/unstructured
data mining technology,
with a patent pending for how it interacts with
numeric/structured data (patent pending) |
|
|
Algorithms
that automate the analysis of multiple dimensions
or factors of data (patents granted and patents
pending) |
|
|
Sophisticated
visualization in 2D and 3D charts and graphs (patents
granted and patents pending) |
OLAP
(On-line Analytical Processing):
OLAP
functionality allows business users to turn raw
transaction-based data into business "insights"
by enabling a multi-dimensional view of their most
important business metrics (such as profit, cost,
time, etc). OLAP presents the numeric data (called
measures) in common business terms and organizes
it into related groups or attributes (called dimensions);
these are groups against which it is important to
understand the business metrics. These dimensions
can also represent a natural hierarchy that enables
the concept of "drilling down" to the next level
of detail. In addition, OLAP architecture delivers
an almost instantaneous response to most queries.
This fast response is essential to maintain analysts'
"train of thought" as they proceed with any particular
thread of inquiry. Having to wait 2 minutes or 2
hours for a query to be processed not only affects
the user's momentum but also jeopardizes the quality
of their analysis. Other standard OLAP features
include summarization, sorting, and filtering of
data to allow the analysis to be targeted to specific
areas of interest. PolyVista has been designed as
a best-of-breed OLAP front-end application that
leverages the performance and feature set of the
Microsoft OLAP database engine (Analysis Services).
|
|
 |
Text/Unstructured
Data Mining Techniques:
PolyVista®
patent-pending "Text Analysis and Mining" module delivers
a complete business solution. We have packaged all of
the necessary processing and text mining routines to build
descriptive text clusters. These clusters are then transformed
into standard dimensions in a multi-dimensional architecture
such as Microsoft Analysis Services. Bringing the structured
and unstructured data together into a common architecture
(data cube) is really just the starting point for analysis.
To complete the job, PolyVista's Discovery Client provides
the perfect interactive analysis interface for the end-user.
PolyVista uniquely delivers the "last mile" by providing
a multi-function analysis tool, combining OLAP browsing,
2D and 3D visualizations, and Discovery algorithms. The
Discovery Client automates the search for new and valuable
business insight.
Additional information on PolyVista’s patent-pending technologies
for Text Mining can be found here.
Embedded
Data Mining algorithms: These algorithms "automate"
the analysis of multi-dimensional data and highlight interesting
relationships for the user. PolyVista software not only
enhances standard OLAP analysis methods, but also delivers
powerful new techniques to help users discover and gain
valuable business insight that would otherwise have gone
unnoticed. We have optimized data mining algorithms to
exploit the performance, aggregations and dimensionality
of these OLAP engines. Our software architecture has been
specifically designed to be extensible and to enable new
algorithms or methods to be quickly implemented. Numerous
algorithms are available that perform statistical searches
for trends and anomalies, while others are targeted for
specific pattern classification or prediction.
What's more, PolyVista uniquely integrates the results
of data mining functions such that these results (called
nuggets or 'insight gems') can be directly manipulated,
validated or enhanced by the user in our OLAP framework.
This capability to take a mined result and use it as the
basis for continued manual analysis illustrates PolyVista's
Interactive Knowledge Discovery (IKD) process. This process
generally includes automated mining to provide an initial
set of "interesting" results (nuggets), from which the
user can select a nugget to manually verify, or extend
the analysis and subsequently tune the mining algorithms
to search for additional relationships.
Available
algorithms include:
|
|
Deviation
- Discovers statistically interesting
data anomalies and ranks the result list for the
user. For example, "What multi-dimensional queries
return cell values where profit is significantly
higher or lower than expected?" |
|
|
Summary
- Discovers the
combinations of dimensions that yield the greatest
or the least measure values. For example, "What
dimensions yield the largest profit when considering
1 or 2 or 3 or n dimensions at a time?" |
|
|
Market Basket
Rules - For a given set of transactions,
what events occur at the same time? A classical
example would be "What products sell well together?"
|
|
|
Association
Rules - Discover a set of multi-dimensional
rules that can be extracted from the data and their
corresponding support and confidence metrics. For
example, "IF customers have a high-school education
THEN it can be said with 93% confidence and 20%
support that they make between $30k - $50k a year." |
| |
|
Decision
Rules - These rules are generated by
a decision tree algorithm. One dependent dimension
or member can be classified by several independent
dimensions or members. |
| |
|
Probability
Rules - These rules are similar in form
and setup to Decision Rules but, here, all selected
dimensions will be part of the resulting rule sets. |
| |
|
Text Mining
- Text clustering algorithms distill
and extract key concepts and issues from free-form
text fields. The categorize results are then merged
together with standard OLAP data, facilitating PolyVista
Discovery analysis of both structured and unstructured
data. |
Fully
interactive 2D/3D Data Visualizations:
PolyVista's visualization capabilities help transform
tables of numbers into shapes, colors and motion to leverage
a users' cognitive abilities to detect trends and relationships
visually. PolyVista's visualizations include more than
thirty chart/graph types, and in addition, fully interactive
2D/3D visual models as well (patent pending). These new
2D and 3D data visualizations are fully synchronized with
the OLAP table control so the slice-n-dice results can
be rendered simultaneously in both the tables and visualizations.
PolyVista 3D objects are fully interactive and can be
positioned to any viewing perspective via the mouse. An
"Animation Play" control enables data sequencing (by time
or any other parameter) to animate the tables, charts
and visualizations, thereby assisting with the identification
of patterns and trends. Samples of PolyVista's 2D/3D visualizations
can be seen by following this link.
PolyVista Solution Architecture
The typical implementation architecture for the PolyVista
solution would include Microsoft's Analysis Services (OLAP
multidimensional database) as the data repository with
PolyVista providing the robust analytics. The OLAP database
could be populated from any available enterprise data
source. Such sources include existing data warehouses
or data marts, text files, spreadsheets, relational databases
or any other data source that supports Microsoft's ODBC
or OLE DB data interface standards. Thus, the PolyVista
architecture is very flexible and extensible, allowing
the capture of data from virtually any data source in
the enterprise for closed loop analysis. |
