P.S.V.

SURYA NARAYANA MURTHY

H.No: 20-119/2, Mob: 9966123977
Dayanand Nagar, Res: 040-27054002
Malkajgiri, EmailID:
[email protected]
Hyderabad - 500047

Summary:
• Expert Data Miner; BASE SAS Certified Professional
• Currently developing Risk Mitigation Models in Basel 2 on Credit Risk,
Operational Risk and Market Risk using Loss Distribution Approach &
Anti Money Laundering applications for BFS domain.
• Working with HSBC Global Technologies Ltd., Hyderabad as Senior
Consultant Specialist-Analytics.
• Worked with SAS Research & Development as Software Specialist-
Analytics & Data Mining
• Worked as an Assistant Manager-Analytics in HCL Technologies Ltd.
• Authored several articles in the field of Statistics and Information
Technology
• Working on BASE SAS 9.1.3, SAS RISK DIMENSIONS, SAS/STAT,
SAS/Marketing Optimization, SAS Marketing Research, SAS Text Miner,
SAS Enterprise Miner, SAS Management Console, SAS OLAP Cube
Studio, SPSS 15.0, SPSS Clementine 10.0 and SPSS Text Analysis
• Used MIS Reporting tools like SAS Information Map Studio, Web
Report Studio
• Key Product Developer for SAS Marketing Automation SaaS, a
Campaign Management tool (Yet to be released) and a user of SAS
Marketing Optimization.
• Experienced in Dynamic Risk Factor Modeling using SAS Risk
Environment.

Work Experience

Working with HSBC Global Technologies, Hyderabad as Senior Consultant

Specialist-Analytics & Data Mining.
Responsibilities include developing data mining applications around Anti Money
Laundering Solutions, Training and Solution Development.
Previous Experience

Worked with SAS R & D, a division of SAS Institute, India, as Software

Specialist-Analytics & Data Mining since August 2007 to July 2009.
Responsibilities include: Developing Data Mining Solutions /
Applications, Project Management, POC design (Proof of Concept) and
Technical Consulting / Training.

Worked with HCL Technologies Ltd., Data Analytics Dept., Noida during
Aug 2006 to Aug 2007, as an Assistant Manager-Data Mining.

Worked in ICFA, Research & Analytics ., as Research Associate / Faculty-

Data Mining for 3.8 years during Feb 2002 to July 2006.

Worked in Jeevan Softech Ltd., as a System Administrator, in

troubleshooting the Windows 2000 Server network for 6 months during
Oct 2001 to Jan 2002. Worked on network utilities and services including
DHCP, DNS etc.,

Technical Expertise and Data Analysis capabilities

SAS Risk Dimensions expert to develop risk management applications in

BASEL II & SOLVENCY II.
SAS Enterprise Miner 5.2 for Predictive Modeling including Decision
Trees, Neural Network.
BASE SAS 9.1.3 programming including PROC REG, PROC DISCRIM and
PROC SQL.
SAS Marketing Optimization 4.4 for optimizing campaigns and mapping
product-customer offers.
SAS Marketing Research for building Conjoint Analysis,
Multidimensional Scaling for developing Utility Functions and Market
Share Simulations.
SAS Management Console for administering various SAS resources and
Marketing Optimization Add-ins.
SAS OLAP CUBE Studio 9.1.3 to develop Multidimensional OLAP Cubes.
SAS STAT STUDIO 3.1 for developing, deploying Multivariate analytical
Models.
SAS Forecast Studio for developing Autoregressive Models and using
High Performance Forecasting procedures for selecting best method for
anomaly detection, forecasting time series.
SAS Text Miner Client for mining textual and verbatim responses and
developing SVDs and Roll Up Terms for textual terms summarization.
SPSS 15.0 Advanced Regression Models and other Multivariate
Techniques.
SPSS Text Analysis for surveys for analyzing verbatim responses.
SAS/INSIGHT, Interactive Data Analysis, in developing regression models.
Windows 2000 Server troubleshooting and System Administration.
Configuration of DHCP Server, DNS Server and Active Directory.
MS-SQL Server Database Administration, including recovery, database
design, Tuning.
Working with CMIE Databases like Prowess and Business Beacon for
extracting large amount of data for analysis.
Working with Intellectual databases like ScienceDirect, MarketLine,
Economic Intelligence Unit, WDI and Reuters Business Insight.

Papers, Articles and Case Studies Published

1. Detecting frauds in Financial Institutions using SAS, being accepted and

published in The Accounting World, a magazine of ICFAI.
2. Detecting financial frauds and number discrepancies using Benford’s
Law.
3. Doing Business Intelligence with SAS 9.1.3, being published in The
Global CEO, a magazine of ICFAI.
4. Returns on Marketing Investments, a research paper published in IIM
Ahmedabad Journal. The paper talks about the returns that SAS Inc.,
realized by investing in the field of Pharmaceuticals.
5. Returns on Marketing Investments, a second research paper
published in IIM Ahmedabad Journal. This papers models a Marketing
Intelligent Solution using which an organization can gauge its marketing
returns using SAS 9.1.3. This paper takes CLTV in to consideration.
Note: Both the above papers are presented and published, at IIM
Ahmedabad conference, held at IIM-Ahmedabad campus, in Jan’06.
6. Adidas acquisition of Reebok, a case study presented at IMT
Ghaziabad International Case Study Conference held from 19 to 20th
December, 2005.
7. Business of Software, a Book Summary, published in BEST BOOKS.
8. An article titled “Business Intelligence for CEOs”, published in ICFAI
Magazine.
DATA MINING Projects and Assignments handled

1. Developing Anti-Money Laundering Solutions:

 Nature of the project: Solution Development
Description: Banks & other financial industries have been heavily victimized
by Money Laundering activities lately. In order to identify suspicious activities
by various customers an attempt has been made to develop data mining /text
mining applications to extract business intelligence from investigation logs, a
framework of comments entered by investigators on various customers who
do a transactional activity over a period of time. These comments are text
mined for detecting patterns and sufficient insights are extracted to prove
whether or not the customer is suspicious and to be reported to the authorities
or not.

Primary role includes application development to compute probability to

launder money and other financial instruments using SAS Enterprise Miner,
SAS BASE and SAS STAT. Besides textual SAR reports were also mined
using SAS Text Miner application.

2. Marketing Automation, Software as a Service:

Nature of the project: New Product Development
Description: Marketing Automation involves creating, managing and
administering marketing campaigns with the power of analytics. With the
intention to offer this product / service as SaaS we ventured in to developing
the product using standard SAS Software: BASE SAS 9.2, SAS STAT, and
SAS Web Report Studio. My role is to develop unsupervised segmentation
using Clustering procedures mimicking the procedures of SAS Enterprise
Miner. Some of the procedures used are PROC DMVQ which is a set of vector
quantization method used to cluster a group of people or entities. The
segments of customers would then be used as input to Marketing Automation
components like Campaign Studio as a class variable to understand the
segment of customer through a granular approach.

3. Genetic Intelligence & Social Network Analysis:

Nature of the project: New Product Development
Description: This effort started with my own IP developing a new product
that can reveal social network ties and their effect on any contagious activity
like churn or fraud. Using SAS IML, SAS BASE, SAS STAT the aim was to
develop a solution that can answer the question of hidden network
relationships between various people and how can this improve the validity of
the model. This effort also lead to develop a genetic intelligence model that
can help us determine the effect of hereditary disease and how it can affect
others in the network apart from family ties. As part of this we had to use
SAS / GENETIC procedures like SAS GENESELECT and PROC ALLELE.
4. Social Network Analysis: Identifying Churners through Network
Visualization Workshop
Client: POC
Nature of the project: POC
Project Description: Visual BI is a hackneyed term in the field of analytics.
With increasing churn rate our client was interested to visually explore the
network groups that exist in their database who do a collective activity based
on the social ties they have between them. With SAS NVW (Network
Visualization Workshop) we created the analytic ready data sets that would
give insights in to the groups of users who commit a similar behavior.

5. Optimizing LTV, Churn Rate using SAS Marketing Optimization 5.1

Nature of the project: Offsite
Client: Qwest Inc.,
Project Description: The effort of computing optimized life time value and
churn rate and thereby assigning product treatments with constraints in issue
is handled in this assignment using SAS Marketing Optimization 4.4. The effort
also lead to preparing a road map for SAS MO from 4.4 to 5.1. Various inputs
involved the contact cost per customer for various channel types, campaign
cost, tactic response, response, churn, margin, tenure of various products
which were used to compute LTV. The challenged involved was to introduce
tactic response scores like loyalty campaign score, gift card response etc in
the MO environment.

6. Global Sequencing and Frequency Analysis

Nature of the project: Offsite
Client: Qwest Inc.,
Project Description: As a consecutive to the customer characterization and
identifying the effect of treatment types of various products to the customers,
our assignment included, estimating churn rate of customer for respective
product types and methods to deal with it considering some potential
predictors. This exercise would lead us in designing a survival analysis model
for product type given the product tenure for individual customers. Key
analytic solutions include: SAS BASE 9.1.3, SAS EM Client and SAS EG.

7. Qwest Customer characterization

Nature of the project: Offsite
Client: Qwest Inc.,
Project Description: Understanding and characterizing customers was the
task Qwest had in hand. Customers across US were treated with various
product features either through TM (Telemarketing) or DM (Direct Marketing)
or EM (Email Marketing) or other types of channels. Our assignment was to
answer key questions relating to the success of a particular channel in getting
most response from customers, and identify feasible solutions in terms of
average number of DM contacts or TM contacts aimed at a customer before he
 or she opts to a DNS list (DO NOT SOLICIT). Key analytic solutions include:
BASE SAS 9.1.3, SAS EG and SAS Enterprise Client.

8. Data Validation for effective segmentation and life time value

evaluation
Nature of the project: Offsite
Client: Qwest Inc.,
Project Description: Qwest is a leader in Telecommunication domain in US
and has roots in other parts of the world too. The objective of this assignment
is to understand Qwest’s customer base, segment the base to effectively
evaluate LTV. We have tracked about 370 variables, that carry customer’s
information, describing their product preferences, product attributes, revenue
generated by this base over a period of time etc., Using Enterprise Miner,
Imputation, PCA, SOM & Clustering Nodes, we have identified profiles of
the segments, that minimize the task of evaluating length of customer
relationship.

9. Data Quality validation & alert generation using OLAP Cube Studio
Client: Time Warner Cables Inc.,
Nature of the project: Offsite
Project description: TWC Inc., is interested to segment their customer base,
where in they deal more with B2B business. To this effect, they were keen to
know the effectiveness of data collection instruments they have deployed, via
surveys, secondary collection methods etc., Our aim in this project is to use
BASE SAS & OLAP Cube Studio to generate fields where alerts are
generated, alert reasons and value of the field for which the alert is
generated. These three generated fields are shown in OLAP Cubes to enable
drill down informing about the fields which require more concern.

10. Evaluating verbatim responses with Telecom giant

Client: Nortel Inc., USA
Nature of the project: External (Offsite)
Project Overview: With customers being surveyed on various textual
variables, and log records maintained in
note log, phone log etc., we intended to extract some key terms that might
help us cluster customers and their
activities and map the same to their case priorities. The sample size was about
2700 observations giving us
hierarchical clusters enabling us naming the clusters with defined
nomenclature.

11. EDCSM, an Event Driven Customer Satisfaction Model

Client: HCL Technologies Ltd., BPO Services,
Nature of the project: Internal project
Project Overview: HCL Technologies Ltd., BPO Services presently handles
huge business account of an UK
 Telecom giant. Recently trend shows a high dissatisfaction rate of its customers
because of factors in call
center efficiency. By including many factors that affect advisors behavior and
like a model was developed in
order to cluster variables which need immediate attention and opportunities for
development. Three models
were developed including Decision Tree (to develop some induction
rules), Cluster Analysis (to develop
clusters of variables) and Measures of Association for model
validation.

12. Evaluating consumer preferences: A Metric Conjoint Analysis

Approach

Project Overview: Conjoint Analysis is used to evaluate consumer

preferences. Given the exhaustive attribute set for an item / category in a retail
store our model answered questions like “how important is each attribute of the
item to consumers?”, “which existing items do consumers prefer?”, “what
combination of item attributes do consumers prefer the most?” and “how well
will item do in the current market”. Several factors were considered in the
model namely: brand name, whether or not a private label brand, price of the
item, perishable or non-perishable item etc., Several attribute levels were
considered in the model. The model delivered output datasets of attribute
importance, combinations of important levels of various attributes taken
together, utility graphs and market share simulation (which gives the simulated
market share).

13. Association Discovery analysis

Client: Safeway Inc.,
Project Overview: With retail operations across UK & US and with
increasing customer base, a problem of identifying consumer’s purchase
behavior had to be answered. With around 80,000 items and categories in
the basket of retail perishable & non-perishable items our client needed a
modeling approach. Our model of Association analysis using Enterprise
Miner helped our client efficiently manage product assortment, thus giving
service intelligence and serve customers profitably. With level of support
and confidence % and other such factors as lift, we could find hidden
relations between items purchased, and also perform promotional analysis.

14. Designing WLR Returners Model for British Telecom (RETAIL) Plc.,
 Client: British Telecom Plc, UK.
Nature of the project: On Site project in BT Premises, London.
Project Overview: WLR is the acronym for Wholesale Line Rental. It
refers to the part of BT where the customers have churned from BT Retail
to some potential competitor namely CarPhone Warehouse. Our idea
is to generate three models namely Regression, Neural Networks and
Decision Trees. We generated lift charts and Cumulative Response
Charts with Classification Matrices which predict the customers who
are likely to return to BT. Our project also includes in designing propensity
scores, sorted by percentile, which helps us in profiling potential
customers who can be targeted for attractive offers. The size of
the dataset is around 17 million records.

15. Designing Cable Returners Model for British Telecom (RETAIL) Plc.,
Client: British Telecom Plc.,
Nature of the project: On Site project in BT Premises, London.
Project Overview: Our model wishes to prioritize the win-back activity so
that only the most valuable and or most likely to return customers are
canvassed with win-back offers. This model lists the propensity scores for
all the customers out of which we get the most profitable customers. The
model encompasses around 4, 20,000 customers and their attributes
including their demographics.

16. Worked on Evaluating Returns on Marketing Investments, ROMI of

BASE SAS 9.1.3, in Pharmaceutical Industry. This project consists of
measuring ROMI, as realized by SAS Inc., by marketing and venturing their
Business Intelligence product, in the Pharmaceuticals sector. My key role
includes developing a forecasting model, and regression model.
Project in brief: The project aims to develop a model that quantifies each
of the variables, representing a medium for targeting the customer, that go
into marketing investments. The allocation of marketing budget has to be
decided on the historical returns on marketing investments. This model
quantifies all the variables based on historical data and will give an
approximate projection of the marketing budget for the coming period in
consideration. This model takes as inputs the different investments that
have gone into these media for targeting the prospective consumers as
well as the returns that has been obtained from the respective avenues. It
also takes into consideration whether a new customer was created or an
existing customer was induced to buy more of the products thus increasing
the customer lifetime value.
 Consequently, the model measures the Customer Perceived Value. On the
basis on the research that we have done, we proposed two factors namely:
Marketing Value Chain Beta factor and Marketing Competitive
strategy Beta Factor. We have also used the model developed for
effective decision making for allocating the marketing budget among the
various individual marketing avenues.

17. Forecasting Returns in Marketing Investments: A Regression

Analysis approach
Of all the areas of corporate spending, one of the toughest is that of
marketing. Measuring ROI can help companies determine what
investments are worth continuing and what are not. The common
investment avenues (as our knowledge of current practices are concerned)
are: advertising, consumer promotions (in the form of rebates and
coupons), direct response, celebrity endorsements etc., And these have
direct impact on the fundamental objectives of the company viz.,
 Acquisition of new customers
 Increasing revenue from existing customers
 Retaining existing customers
 Supporting higher margins and prices (by increasing brand
equity)

The goal of our model is to embed a return on investment mindset and

process in to the overall marketing infrastructure. In our model we
measures both qualitative outcomes (customer perceptions), along with
quantitative returns (in the form of sales).

Our regression model includes a good number of factors that directly or

indirectly relate to the complex perception of the customer, including his
own awareness stage. Our model acts as a marketing intelligence
solution, enabling managers and decision makers to evaluate ROMI as a
percentage of Total Marketing Investments, with sales generated
by qualitative and quantitative factors including. Using the model
developed simulations can be done to achieve forecasted sales figures for a
company for a particular amount of investment.

18. Worked on Discriminant Analysis, in classifying depositors and loan

takers as defaulters.
Project in brief: This assignment includes a discriminant analysis
model that we developed in classifying customers in to defaulters and non-
defaulters. By taking a representative sample of 500 customers, out of 700,
we designed a discriminant model which can help us deciding those factors
that categorize customers in to good and bad credit risks. As experienced
bankers, supposed, we came to conclusion that factors including Debt-to-
 income ratio, credit card debt, number of years the customer
stayed in a place of stay and number of years with the current
employer turned out to be best discriminators. The validity of the model
measured in Wilk’s Lambda was less than ‘1’ and so did the WIilk’s
lambda of the variables, thus making the model successful in
classifying the customers in to correct categories with a
probability of correct classification as nearly 80%. We then applied
this model to the remaining 200 customers and the model seems to be
significant.

Qualification
Masters Degree in Computer Applications, MCA from Osmania
University.
Bachelors Degree in Statistics and Computers from Osmania University.

Certifications
BASE SAS 9.1.3 Certified Professional.
SAS® training and participant certificates in SAS Enterprise Miner, BASE
SAS 9.1.3 and SAS Time Series Forecasting.

Personal information

Father’s Name : P. Satya Narayana Murthy

Native : Indian
Languages : English, Telugu, Hindi.
Address : H.No:20-119/2, Dayanand Nagar, Malkajgri, Hyderabad-
500047.

Date:
Place: Hyderabad P.S.V.Surya Narayana Murthy