Bank data analytics

-- particles
Situation

Build fast and reliable platform capable to collect, process a vast amount of data from many sources to make analytics assisting financial decisions. To successfully handle continuously growing amount of data from various sources a flexible, fault-tolerant and robust tool is required. A data warehouse solution should be both capable of loading data from numerous complementary systems and providing data-marts and extracts for BI and analytic tools and do it quickly. Taking into consideration the throughput of customer-related data, DWH and multichannel data collection might appear to become a bottle-neck.
Data analysis and aggregation play basically the most important role in solution architecture. On this level, client scoring and credit limit are calculated based on information given by the customer, from credit histories’ bureaus, social networks, and 3rd party databases. All the magic should be done in real-time for a dozen seconds. Moreover, operational fraud detection is a “must have” feature in remote banking. And classical big data questions had to be answered: how to clean, normalize, de-duplicate, enrich data, basically how to bring value to data.

Solution

To meet requirements, it was needed a time-proven workhorse customer relationship management solution and with no doubts that were an Oracle Siebel CRM” Head of ITC Competence Center. Since now it is possible to process the full cycle of executing the contract from the client’s agreement to issuing a card from a Siebel screen of customer manager for less than 30 minutes. Such an impressive result became possible due to the seamless integration of CRM, mobile application for mobile bankers, Unity BARS core system, Interactive Intelligence Contact Center Platform, SAS via enterprise service bus WSO2. “Having a lot of systems in maintenance we dealt with the challenge to keep them in touch and function smoothly and sound in integrity. Service bus implementation made it real and also enhanced security level.” CTO Intellica.

Technologies and skills

Technologies

  • Hadoop
  • R
  • Vertica

Solution

  •  HP Vertica
  • Cloudera
  • Tableau
  • Oracle Siebel CRM
Result

As a matter of fact, the team’s determination and resolution led to the competitive edge for business and better client satisfaction. The bank became the first branchless bank in Ukraine. “Starting from greenfield we ended up with 12% negative debt service after the first year. Although, we assume further enhancement through using neural nets and genetic algorithms for risk estimation models for decreasing fraud, even though 12% is still unachievable for “classic” bank.” said Chief Risk Officer of the client’s side.

Service Type

Fully managed Project Delivery

Business domains

Financial Services

Expertise

Data warehouse, mobile application, web application

Project team

1x Project manager

1x Senior Data Scientist

1x BI developer

1x Data Architect

1x Data Management engineer

1x Fraud detection specialist