Financial Services Business Analytics: Evidence-Based Model

This diagram is a representation of an universal evidence-based business analytics model for the financial services industry. This model takes into account data from customers, investments, deposits, and compliance issues. It is flexible so it can adapt to any situation. It is important to notice that predictive modeling and data mining is done in a separate server than the enterprise data warehouse (EDW), otherwise it would slow the EDW to a crawl. On the other hand, this model complements business intelligence reports. This models uses regression analysis, vector analysis, and clustering (partition) analysis. It takes into consideration that data is three-dimensional (3D). It is flexible enough to add association rules and time-series analysis. The front end incorporates depth-analysis in order to leverage human-computer interaction (HCI) at the enterprise-level (i.e., anybody within the enterprise should be able to use it. It should take three months to deploy this model. If you want assistance implementing this model please contact me at Hewlett-Packard: alberto.roldan@hp.com

Evidence-based Enterprise Business Analytics Model: Turning Around in a Quarter

Companies can implement in 90 days a fixed cost and evidence-based decision making systems that would allow them to become flexible in this economic climate. Science and technology has improved to such a degree in the last three years that companies can use off-the-shelf software to do business analytics which used to be reserved to a few highly skilled professionals within an organization.

The diagram above represents an example of the architecture for an enterprise business analytics module in the healthcare and biotechnology industries. This same business analytics module can be used in any industry or company. Small and medium sized businesses can implement this business analytics architecture using an analytics-as-a-service model. Large companies can implement the same model in-house.

Three-Dimensional Business Analytics: How Deep is the Ocean?

Business data, like our universe, is three-dimensional. Nevertheless, business analytics tends to be flat or two-dimensional like an Excel table or chart. The difference between a two-dimensional analysis and a three-dimensional analysis is depth. Depth perception allows an individual to accurately determine the distance to an object. It is a characteristic mostly seen in higher species.

In analytics, depth is referred to as dimensional analysis. Dimensional analysis is used in engineering, physics, and chemistry to understand the characteristics of multi-dimensional data. It is used to formulate hypotheses about the data, which are later tested in more detail. In business analytics we can create a three-dimensional variable that allows the end-user to “see the depth” of the data. This variable is called a three-dimensional vector analysis. This variable, when combined with cluster analysis and a visualization tool, answers the proverbial business question: how deep can I go into my data and see patterns in which sound business decisions can be made?

Business Case for Analytics: Explaining Cluster Analysis

The issue for business analytics is how to explain the return of investment by implementing enterprise analytics. One of the mathematical concepts involved in business analytics is cluster analysis. Cluster analysis is the utilization of a mathematical algorithm to group data that has similar characteristics.

Most datasets contain information that is three-dimensional (3D). For example, medical data include multiple values: diagnosis, medical procedures, prescription medicines, tests results, age, gender, and length of stay. The issue with most large datasets is how to analyze and visualize 3D data in a two-dimensional format. Cluster analysis allows the partition or segmentation of 3D data.

The best example that I have found to explain cluster analysis is brain imaging. In the example below, you can three images of the brain. The first image is of the top of the brain, the second of the side, and the third one is from behind. You would use a cluster analysis, to classify what areas of the brain is grey matter, white matter, or fluid. In our brain images, clustering analysis grouped the grey matter as red, the white matter as blue, and the brain fluid as green.







The advantage of cluster analysis is that it allows anyone within a company to make decisions based on a clearer picture of its 3D data.

Explaining the Value of Business Analytics: Clarifying driving factors in a decision support system

There are many ways to use data mining and predictive modeling to find patterns in large volume of data and transform them into actionable information. I have found that a combination of data mining and predictive modeling techniques are necessary to separate the clusters of data and clarify the driving factors. For the purposes of this posting, a vector is a multidimensional variable that uses mathematics to predict the probability of an event. When a vector (or predictive modeling) is used in conjunction with a clustering technique (separate data into groups of similar characteristics) the result is a decision support system that separate the clusters in the data and clarifies the driving factor.

The best visualization to explain this concept is in the area of graphics. The graphic to the left illustrates how a vector can clarify an image using vector and cluster analysis. This methodology can be used in any industry. The predictive modeling science and computer technology allow companies to build this capability efficeintly throughout an enterprise using products like Microsoft SQL 2005 Server Analysis Services, or SAS Enterprise Miner.

Intelligent CRM Requires Business Intelligence

"...enterprises use BI applications that can highlight information that will lead to meaningful change in company performance. Rather than just running reports with reams of information, Herschel recommended using reports with "traffic lights" or other signals that alert the reader to the handful or so of the most meaningful data points." This is suggesting to use a methodology that separates the data (cluster analysis) and clarify the driving factors within those clusters (predictive modeling or scoring the predict probability for an even to occur).

Microsoft Introduces Tool for Avoiding Traffic Jams

Microsoft strategy of integrating data mining, predictive modeling, and artificial intelligence in all its product is what separates them from their competition. The application of these technologies to their product offering would change the landscape of how we conduct business for the next 50 years. Imagine integrating the Clearflow AI with the data mining algorithms of SQL 2005 Server Analysis Services, and adding a powerful visualization tool to a product like Excel...wow!

Information on Some Trends in the Disease Management Industry

A very good summary of the Blues survey pertaining predictive modeling in the Blues Plans.

MARKET SEGMENTATION: A Neural Network Application

This paper is about the utilization of business analytics in the tourism industry. The methodology could be applicable to places like Dubai, or for the cruise industry.

On optimizing the selection of business transformation projects

This is an IBM paper that presents a good methodology for business transformation using predictive modeling and business analytics. This approach could be used in supply-chain cases. From a practical approach, if you have a Microsoft shop, you can use Analysis Services (2005 or 2008) Logistic Regression or Linear Regression and use the predict probability feature to get the multidimensional vector. At the front end, the cluster analysis function of Excel 2007 will help separate the cluster. By adding the vector to the cluster variables this algorithm will give you an interactive or dynamic visualization that will separate the clusters in the data, and clarify the driving factors.

How to compete on analytics: apply it

This SAS.com blog is excellent and it reflects the changes that are happening in the market place: companies are incorporating business analytics as part of their strategic objectives. The SAS Institute is far ahead in their analytics products than their competitors. They have business modules for almost any industry and great dynamic visualization tool. They are the pioneers and leaders in this area.

Business Competency Model: Turning Around in a Quarter

Companies need an organization that allows faster responses to market needs, and differentiated products or services to survive in changing economic times. In the area of business analytics the issue is to leverage the knowledge within an organization that allows faster responses to market needs, and differentiated products of services. Business analytics allows everyone within an organization to visualized potential trends internally and externally. The premise is that the greater number of people within an organization that have access to the entire data the faster changes can take place. Since the amount of data is so large there must be a visualization tool that has the capacity to show the entire dataset, while simultaneously separating the clusters of data (i.e., what is in there?) and clarifying the driving factors (i.e., why is in there?).

Companies that outperform the market in tough times have shown to change their organizational model to reflect the times. In 2001 I wrote a paper which states that the changes in an organization start at the executive management committee level and it filters down to the entire organization. Therefore, companies which want to use business analytics to achieve their strategic objectives must have a business analytics committee within the executive management structure. I like to think about this area in terms of parents setting the examples within their own families: an example of how you do something has a greater impact than speaking about it.

Business Analytics – Getting the Point

This article explains the evolution from business intelligence to business analytics: from getting the data to using the data.

Default Risk, Asset Pricing and Debt Control

In today's world markets this is an important paper. The premise is that you need to analyze a company's credit risk when you evaluate a company's assests. This is relevant because insurance companies, savings and loans, banks, mutual funds, pension funds and other companies have invested in what is considered high-risk investments like mortagage-backed securities. We might find that this credit-risk issue may touch other organizations like insurance companies and pension funds.

DEVELOPING RICH INSIGHTS ON PUBLIC INTERNET FIRM ENTRY AND EXIT BASED ON SURVIVAL ANALYSIS AND DATA VISUALIZATION

The methodology of this paper can be used to determine the "health" of public companies in the area of credit risk management. These visualizations can be done in Excel. I would add a multidimensional vector (or variable) in order to clarify the driving factors.

Seeing is believing: Designing visualizations for managing risk and compliance

This is relevant paper by the IBM Labs in visualization analytics. Although the paper orientation is toward risk management compliance in the Sarbarnes-Oxley area, its general principles have applicability over all analytics. For example:
1. A visualization should be thought of as a user interface to a control task, not as a report or a report component.
2. The patterns that are important for users to manage are at the level of the
controls
3. Visualizations must be able to evolve as the process for an individual or a group evolves or as the overall compliance process evolves

Computer-aided Detection of Lung Cancer from Computed Tomography Images

Good technique to separate the clusters and clarify the driving factors.

Sr .Net Architect

I usually do not do this, but there is a friend that is looking for a Sr .Net Architect in the Nashville, TN area. It can pay up to $120/hr.

Conference or Seminar on Data Mining and Visualization Tools

A member of the Business Analytics Group wants to know if anyone knows a conference or seminar that will cover the issue of data mining and visualization tools. If anyone has any information please let me know.
Alberto

More data usually beats better algorithms

This article pinpoint something that has been true for a long time: more data usually beats better algorithms. Therefore, assuming that the data mining algorithmns are not the issue (assuming good science behind them, which I have found in all the major software vendors), the issue then becomes the quality of the interactive visualization tool that allows end-users to make better decisions. Fed Chairman Bernanke, when at Princeton, published a paper that is complimentary to this issue.

Business Analytics is Evolving Says SAS

Good article explaining the definition and value of business analytics.

Schlegel on Search, Analytics and Visualization

Gartner's Schlegel on next BI industry consolidation trend: predictive analytics and visualization tools.

People would rather live with a problem they cannot solve than accept a solution they cannot understand

This quote by Robert Woolsey is the main business analytics issue. The science and the technology nows allows predictive analytics and data mining to be available to every business regardless of size and complexity of the problem. The issue is how to explain the solution to organizations. I believe that an interactive visualization tool, that mirrors the goals and strategy of an organization, is the key to make businesses understand and embrace the solution. A picture speaks a thousand words!