Bank Insider Interpretation: Everbright Bank How to use large data?

In the big data age, the data is competitive, the Internet financial impact of traditional banking. It's better to listen to people inside the bank about their work, how the banks themselves look at big data and how big data works inside traditional banks.

The following is Everbright bank Liu in the "2013 Cloud World Congress," the keynote speech summary, from the framework to the application of the bank in the big Data age is how to build infrastructure and the use of large data technology:

Everbright Bank's own internal in the 2013 is Everbright Bank's big data year, from data governance, data architecture planning, Data Warehouse continuous upgrade and expansion, application of new technology to data analysis, mining, related technical preparation, and related assistance, we are basically ready to use large data to play value. We would like 3721.html ">2014 years we will have more exploration in this regard."

There are three aspects of data services:

On the one hand, we belong to the National Commercial Bank, the first step is to provide professional data services to the head office; the second aspect is to provide data services to branch offices; Finally, the focus of business development, the continuous provision of data analysis reports, the use of large data to find business linkages, discover business opportunities.

To cite a few examples, the first example is that domestic banks are rarely achievable, and that's what we dream of doing. This is the ABN AMRO Bank, which has a very simple picture on their website, which means that when the customer comes, they will see an advertisement, such as a car loan. Customers see this ad, want to do a calculation on the site, but he mistakenly use the mortgage calculator. The bank immediately sees the user using the mortgage calculator and automatically turns the page ad the user sees into a mortgage ad. This mortgage ad only this one customer can see, all the other customers see is still just the car loan ads.

What does that mean? It shows that the bank is doing real-time marketing on its web-banking portal. Analyze the customer's clicks on his online bank, and analyze the data of non-financial transactions and do the marketing in the future. This is a very interesting place for ABN Amro, and few banks in the country now can do real time marketing.

Everbright Bank recently in the development of community banks, also used in large data analysis technology, we will get a lot of external data, including 58 of the same city data, public comments data, soufun data, as well as some community planning data, including the community banking business model positioning, and differentiated to the Community Bank resource allocation and so on.

The next example is Sina Weibo's public opinion monitoring and analysis. For the Internet such data retrieval, including the use of Chinese semantics for relevance analysis, now many companies are doing, Everbright Bank first to do some exploration. Analysis of the current Everbright Bank in the public opinion of the data, the discovery of financial needs is still the most important, this is the characteristics of everbright.

In addition, in response to changes in customer sentiment, to do some marketing. can also be divided into regions, different regions, different characteristics of the Internet users are not the same.

How do you use great data? We have a description, from the initial exploration to the use of large data, to the end we have to deploy the use of applications, which includes a number of steps, including the acquisition of large data, including how we use large data, in what areas to analyze, including our speaking of customer marketing areas will be used, Eventually we will use mature large data applications, including some of our large data related management system to standardize the use of large data.

From the data architecture point of view, we may need to break through the original data application architecture to do, including the current data requirements faster, just mentioned the request to t+0, now we do is basically t+1.

In terms of architectural design objectives, we need to meet diverse data service requirements, transition from a single platform to a diverse structure, and ultimately a wider variety of data processing requirements. This requires the bank to have an open mind, with more manufacturers to exchange and collaborate, to make the real application of the structure.

Finally, the application of a new technology, that is, Hadoop. While many people are saying that Hadoop does not represent large data, Hadoop is a very important application architecture for large data. Now the bank is basically the structure of large data processing, including we now have the integration of Oracle and so on, lateral extensibility is limited, but for Hadoop, the lateral scalability is very strong, the cost is very low. So the future of Hadoop is very competitive.

What is Everbright's attempt to do with Hadoop? Because the bank has historical data, we from 1999 after the centralization of the system, more than 10 years of data has been in the state of not online, there will be a lot of historical verification requirements, including customers, regulators, prosecutors, will come to our bank for historical data verification. The efficiency of this verification is very low, how to solve the effectiveness of verification support? It's Hadoop.

The historical data query function on the basis of Hadoop, in fact, the most important thing is that we really use this open Hadoop architecture technology to do a very traditional banking business, which is of course, we now see the banks dare to use Hadoop in the business of innovative attempts, we now use pretty good, I think we're going to have more fields to try Hadoop technology. Finally, it is the way to activate the offline assets and activate the data value.