E-Risk Services (DW & App Unification)

Client & Industry

E-Risk Services, a provider in the insurance technology space, needed to modernize its insurance underwriting systems and unlock the potential of its vast data assets for better analysis and decision-making.

The Challenge

E-Risk faced several interconnected challenges related to its existing systems and data infrastructure:

• Multiple, disjointed underwriting applications created operational silos and inefficiencies.
• Existing systems lacked flexibility, scalability, maintainability, and recoverability.
• Inability to easily explore, mine, and analyze data from various sources (structured, unstructured, historical, third-party).
• Need to integrate diverse datasets, including application data, claims data, surveys, and external sources like IRS 990 tax data, for comprehensive analysis and potential fraud detection.
• Lack of a centralized platform for ad-hoc data mining and reporting.

Solution by Nimbletec

Nimbletec provided lead database architecture and senior software engineering services, focusing on both application unification and data modernization:

• Review & Strategize Underwriting System Modernization: Conducted in-depth reviews of existing insurance underwriting systems (developed potentially using ASP.NET and SQL Server) to uncover critical improvement opportunities in flexibility, scalability, maintainability, availability, and recoverability.
• Unify Disjoint Underwriting Applications: Architected and developed core software abstractions for product types and aggregated interfaces, relocated common logic into shared libraries, grouped products into logical classes for uniform processing, developed a generic execution strategy for loading/processing applications by product type, and repackaged legacy models into a cohesive, unified platform. Development managed within Microsoft Visual Studio and Eclipse, version controlled with Team Foundation Server (TFS).
• Construct Cloudera-based Big Data Warehouse: Established and configured a Cloudera distribution of Apache Hadoop, creating a flexible Data Warehouse (DW) "playground" environment hosted on Amazon Web Services (AWS), specifically optimized for data exploration and the development of advanced data mining tools.
• Develop Big Data ETL/ELT Pipelines: Engineered robust Extract, Transform, Load (ETL) and Extract, Load, Transform (ELT) processes to cleanse and load large volumes of unstructured data into the Hadoop Distributed File System (HDFS) and structured data into queryable formats using Apache Hive and Apache Impala for experimentation and analysis.
• Construct High-Performance Spark Processing Jobs: Built efficient, high-performance data processing jobs using Apache Spark, leveraging both Scala and Python programming languages, to analyze diverse and complex datasets including Internal Revenue Service (IRS) 990 tax forms, insurance claims, customer surveys, and historical application data.
• Integrate Diverse Data for Advanced Mining & Fraud Detection: Unified historical application and claims data from SQL Server with various third-party data sources within the Cloudera Data Warehouse to power sophisticated data mining initiatives and develop capabilities for potential fraud detection.
• Engineer Custom Reporting & Ad-Hoc Analysis Tooling: Developed specialized data extraction jobs using SQL Server Integration Services (SSIS) to pull data from legacy SQL Server databases and unstructured document sources; created bespoke data extraction and aggregation tools to facilitate ad-hoc data mining and reporting via QlikView dashboards and SQL Server Reporting Services (SSRS).

Key Results & Impact

Nimbletec's strategic application unification and data modernization approach yielded significant improvements for E-Risk Services:

• Unified Underwriting Platform & Improved Efficiency: Provided a single, cohesive point of access for managing and tracking insurance products, significantly improving operational efficiency and reducing redundancy.
• Enhanced System Architecture & Future-Readiness: Increased the flexibility, scalability, maintainability, and recoverability of core underwriting systems, positioning E-Risk for future growth and adaptation.
• Enabled Powerful Data Exploration & Experimentation: The Cloudera Data Warehouse playground empowered analysts to experiment with and derive novel insights from diverse, previously siloed datasets using tools like Hive and Impala.
• Facilitated Advanced Analytics & Fraud Detection: Integration of various internal and external data sources within the Hadoop ecosystem enabled sophisticated data mining using Spark (Scala/Python) and laid the groundwork for advanced fraud detection capabilities.
• Improved Reporting & Ad-Hoc Analysis: Enabled faster, more flexible ad-hoc reporting and data visualization through custom ETL/ELT pipelines feeding tools like QlikView and SQL Server Reporting Services (SSRS).