Client         : Financial Client
Project       : Analysis of retro-scores for risk assessment
Objective   : To identify the risk associated with customers for the credit card products.
Method     : Identify propensity scores for every customer in order to determine his or her potential non-repayment capacity (risk).

Process:
o Using the historical data of credit card transactions a logistic regression model has been developed to identify logit scores for every customer.
o Then this data is compared to the data derived from different credit bureaus.
o Based upon the scores, risk is assessed for every customer.
o Which is being used to predict the future behavior of customer towards credit card products and customers are profiled in to different segments for the    decision making purpose

Tools used: SAS9.3.1, TeraData, Unix, SQL
Frequency: Monthly

_________________________________________________________________________________________________________________________________________________________
Client         : Banking
Project       : Online targeted advertisements
Objective   : To identify and target online customers for online advertisements (products)
Method     : Based on the probability scores, potential online customers are identified and targeted

Process:
o Based upon a few months of historical data a statistical model has been developed to predict the probability scores for every customer for different products.
o Then those scores are used to identify potential online customers for each and every product that different lines of businesses request.
o Some customers may be qualified for more than one product. Then the tie is resolved using some in built management criterion.
o Then the identified customers are targeted online for different products.
o Campaign is run every week and ads are launched from a predestined golive date.
o Then the customer's click information is collected to assess the effectiveness of the campaign.

Tools used: SAS9.3.1 WIN/UNIX, TeraData
Frequency: Weekly