Bharat Thakkar

Credit Bureau

Building a Scalable, Multi-Country Credit Bureau Platform

Designed and implemented a large-scale, customizable credit bureau platform being offered in multiple countries to operate with financial institutions and regulatory bodies.
The system ingests large volume of data from diverse financial institutions, performs advanced identity matching, and generates unified consumer credit reports for banks.

Credit Report generated from Credit Bureau System

The Challenge

  • Aggregate data from multiple banks and lenders.
  • Handle massive volumes of consumer credit data.
  • Resolve duplicate and fragmented identities.
  • Provide a single unified credit view per consumer.
  • Be configurable across countries.
  • Deliver reliable reporting.

The Solution

  • Implement Data Ingestion Layer.​
  • Perform Data Validation.
  • Perform ETL & Data Processing.
  • Build Identity Matching Engine.​
  • Rule based consumer identification.​
  • Implement Data Storage
  • Implement Report Generation
  • Deliver reports on UI and Webservices.

Key Outcomes

  • Unified consumer credit profiles across institutions​
  • High-volume data processing.
  • Customizable across multiple countries.
  • Rule based engine to match consumers.​
  • Improved data accuracies.​
  • Customizable report for different clients.

Implementation Process​

Data Ingestion Layer​

Data received from external banks via batch files.
Supported multiple formats and schemas
Designed for high-volume, incremental ingestion.

ETL & Data Processing

Implemented ETL process.
Data validation & cleansing.
Data Transformation.
Data Standardization and matching.
Data caching.

Name & Address Standardization​

Third party integration for name and address standardization.
Unique Name and Address parsing based on country norms.
Consumer and Business name and address differentiation.

Identity Matching Engine​

Advanced Identity matching algorithm
Identified consumer using Name, Address, ID numbers & Phone numbers.
Proprietary rules to ensure unique consumer identification.

Data Storage Layer

IBM DB2
Permanent storage of records
Structured, relational credit data.
MongoDB
Consumer data stored as documents
Faster retrieval for data validation.

Credit Report Application​

User Interface for accessing consumer reports.
Enabled consumers search using name, address, ID and/or phone numbers.
Match the consumers and pull together accounts, loans and other data.

Technical Stack

  • Java, Java EE
  • Pentaho ETL
  • Mongo DB & DB2
  • IdentIQ / SAS
  • JBoss webserver