RSS
Логотип
Баннер в шапке 1
Баннер в шапке 2
Project

VTB 24 creates storage on the basis of Data Lake for the analysis of client data

Customers: VTB 24

Moscow; Financial services, investments and audit

Product: Projects of DWH
Second product: Apache Hadoop

Project date: 2017/01  - 2017/05

On August 7, 2017 VTB 24 announced creation of the Data Research Platform platform for storage and a research of client data on the basis of Data Lake ("the lake of data" — the term describing any large amounts of data. Actually it is a repository in which the huge volume of "crude" data is stored in their initial format until they are not used).

Project Progress

By a preliminary estimate, model on the basis of data on a customer behavior are capable to bring to bank annually to 1.2 billion rubles. additional income and 30 billion rubles of additional sales.

According to VTB 24, Data Research Platform allows bank to analyze client behavior in details. In it all information on borrowers which is used further in work of the interested banking systems accumulates. The forecasts of behavior received as a result of researches will allow to create the optimal credit conditions and individual interest rates based on risk assessment on a credit product.

The assigned task in world practice is implemented through creation of Data Lake and creation on its basis of the research platform of data processing which are loaded on it without any conversions. For work with client information means of the deep analysis of Data mining will be used.

File:Aquote1.png
In 2016 we implemented the large project on processing and the analysis of client information. Result of its implementation — significant increase in the key financial performance of VTB 24 due to improvement of client offers. And a system analyzes only 60% of all which are available in a databank. During creation of Data Researh Platform we cover 100% of data and we check our hypotheses of client behavior and its influence on solvency. The analysis of this information will allow to reduce risks when crediting and to improve conditions of bank on credit products — the board member of VTB 24 Alexander Sokolov told.
File:Aquote2.png

File:Aquote1.png
We are interested in integration of IT into business. When choosing the innovation processes we are guided by those projects which implementation will provide the maximum effect. Creation of the platform according to the analysis of client data using Data Lake will allow us to provide not only storage and processing of a huge array of historical data, but also to reach solving of tasks of analytical CRM, risk management new level, to receive examination in deep data analysis, to grow up own scientists and users of the Data Architect and Data Scientist profiles" — Sergey Rusanov, the board member, the director of the department of banking and information technologies of VTB 24 added.
File:Aquote2.png

Results of the first stage

For August 7, 2017 VTB 24 completed the first stage of works on setup, piloting of mechanisms, approaches and technologies and as the result, unrolled the Data Researh Platform platform. Now it ensures functioning of users of the Data Architect and Data Scientist profile with complete security of data access.

At the same time, the test environment is unrolled, disaster tolerance of the solution is provided. Data Research Platform is implemented in Hadoop ecosystem using Open Source-технологий and industrial enterprise-solutions.

Implementation of the improved models taking into account enrichment by "crude data" and reaching the planned business effect will become the next stage.