Big Data: the first results

Olga Gorchinskaya: Yes, a certain activity concerning big data, in-memory computing (calculations in RAM) and the accompanying technologies in the market is, but after all I had an impression that the interest of customers in such technologies and tools is at the level of experiments so far. Even in banking sector and a telecom where already began to work with Big Data, it is, as a rule, about the stage preceding industrial implementation.

So it is not necessary to speak about mass commercial demand for such technologies so far, but the set of research projects, and not only with the assistance of system integrators, but also forces of internal IT teams of customers is carried out. For example, the VimpelCom operator in many new IT projects uses the instruments of performance improvement connected with Big Data, including Hadoop and NoSQL of the database and receives excellent results. In Sberbank pilot projects for assessment of advantages which can give Big Data are carried out. But real deployment of such tools so far only in plans. There are also exceptions in the market when the companies really use all power of big data, for example, it is Tinkoff Credit Systems bank. It is clear, that such projects very famous for all market.

TAdviser: If to speak about the analysis of Big Data whether correctly we understand that first of all financial institutions, banks became interested in it?

Olga Gorchinskaya: Yes, it so, because the companies of these industries most zrela in data analysis. Most of them already implemented standard BI tools and even profound analytics. For example, practically in all large banks not only tools for the interactive analysis and creation of dashboards are implemented, but also mathematical and statistical techniques in risk management departments therefore it is quite natural to move further in the direction of mastering of new analytical means are used. At the same time it is impossible to tell that Big Data need to be used only for new business challenges. Very often they allow to look in a new way at already existing tasks, to offer their new treatment and methods of the solution.

For example, in credit departments of banks the assessment of profiles of clients, their reliability, solvency (credit-skorring), including means of data mining is traditionally carried out. Use of Big Data, for example, allows to add to this task a new edge due to expansion of structure of data on clients which can be received from social networks. Similar projects go in several banks, for example, in Home Credit, Tinkoff Credit systems, Sberbank now.

Certainly, potentially very serious demand for big data and profound analytics is expected also in a public sector, but here initiatives are implemented more slowly. Technologies of Big Data, calculations in RAM in the state organizations will be demanded also because there as work hard with unstructured information anywhere. FORS has many customers in a public sector. I can cite as an example the project in the Fund of Compulsory Health Insurance (FCHI) where we implement in-memory technologies and profound analytics now, upgrading the existing analytical system. More than a half of revenue of FORS from the BI direction is the share of a public sector, and we expect that this proportion will remain.

'TAdviser: The FORS in examination in the field of a business intelligence relies on a line of Oracle, it is the whole range of products, and not only program, but also hardware-software. Including it is the specialized systems for planning and budgeting. Tell, please, this direction is how demanded? How does it in general develop?

Olga Gorchinskaya: Systems on budgeting and planning of Oracle are introduced into the BPM/EPM block (enterprise performance management) where also solutions for consolidation of the reporting and strategic planning belong. The most popular of them it is, of course, planning and budgeting systems. We note stable interest in the systems of the class BPM. These are very clear solutions for customers, the effect of which implementations is easily estimated even quantitatively. If the effect of implementation of BI can be postponed, long-term, difficult vyrazimy in the form of any specific indicators, then with BPM everything is simple: for example, the planning cycle is several times reduced, and it is already very essential economic effect.

Therefore this class of systems also draws such attention. Planning and budgeting in most the companies is anyway conducted, at least at the level of Microsoft Excel, and EPM platforms like Oracle Hyperion Planning is already following, more serious level. Most the large companies implemented such platforms, now we note interest in them from medium business, and not only in Russia, but also in the CIS where we keep projects too. Quite often such projects gain additional development as in the large organizations in EPM systems works hard users, very difficult budgeting models are used, and performance becomes a big problem. In this case additional solutions such as, for example, in-memory analytics technologies from Oracle company which are built in the hardware and software system Oracle Exalytics are required. The FORS has a strong team on implementation of Oracle EPM including in the direction of performance improvement of the existing systems, we develop this examination very actively.

TAdviser: Experts of Gartner say that one of important trends of BI for the next years is a development of analytical platforms, including regarding interfaces, for benefit of such user-friendly, friendliness to end users to whom tools for easy surfing of data, including unstructured type, drawing up independent requests are offered. Oracle proposes some solutions of such plan? Are available and whether they are implemented in Russia?

Olga Gorchinskaya: As soon as BI systems appeared as a class, experts began to say at once that they should be very friendly to users: all instruments, means of drawing up requests should be simple and clear, and an ideal system should allow business users to formulate them independently. Actually, all modern BI platforms already very close approached implementation of this concept. Besides, also business users in many respects changed. That abstract manager whom the interface of a business intelligence is expected, today already experienced Internet surfer who with ease is able to state requests to search web machines, to find necessary in network, to understand interfaces new and not always successfully cut out websites.

Therefore, in my opinion, not so much simplicity of the interface, how many understanding by the user is important now that to it would be important to be known, and, of course, his motivation to search and the analysis. Therefore BI vendors work on support of opportunities of "active analytics" and intuitive search and many already provide such means. I will give an example: let's allow, profit fell in the company what the head should find in an analytical system to understand the reason of it? To modern BI platforms, including in Oracle BI, the set of ready scenarios on such case "is sewn up". For example, having clicked profit digit, the head right there will receive two new - stock balance and the number of new requests from clients and will be able to compare them. The situation can consist not only in a lack of demand, but also a lack of products that to satisfy it. A system automatically issues information which is not the answer to accurately formulated question, and helps to understand the reason of this or that event more likely.

TAdviser: But Big Data are such array of information where we do not foreknow either its structure, or attributes of future search?

Olga Gorchinskaya: Approximately so. There is absolutely a different approach, working with such arrays of information, we often not absolutely precisely understand that it will be useful to the solution of a business challenge. Exactly here possibilities of intuitive search and the analysis which can push to formation at the user of new hypotheses, quickly allow to check them therefore gradually becomes clear are very demanded how to solve the problem existing business. Such products in the market appear, for example - Oracle Endeca. This solution is already available in Russia, and we do different demonstration industrial solutions on the basis of this platform, integrating it with instruments of statistical investigations, data mining and means of linguistic processing of texts.

We have big and very interesting project for ROSEKO company, the member of the Russian Society of Appraisers (RSA) where the analysis of the Big Data about real estate value taken from the Internet allowed to construct very exact, tied to the map model of assessment of a real estate, including for the state needs which can be used further across all Russia already in "cloud". This solution works at Oracle Endeca, the instrumental environment of statistical investigations Oracle R Enterprise and linguistic products of RCO company is also involved in the project.

TAdviser: Let's dispel the myth that the "heavy" artillery of BI is necessary only to banks, a telecom, retail. The FORS, for example, proposes analytical solutions including for the sports (!) organizations. Than still you can surprise us?

Olga Gorchinskaya: The business analysis is necessary, certainly, not only to the large companies, but also medium business up to small and even to individual entrepreneurs. Banks, retail, a telecom traditionally in vanguard of BI automation, they to it are pushed by both the competition, and great opportunities for investments into IT, but all other industries also move in this direction. BI solutions can represent both powerful platforms, and small the narrow ground systems for specific objectives. Of course, it is interesting to FORS to move not only on the blazed road, but also to constantly do something new.

For us the analytics for sports clubs, first of all football became such new area, but also not only. We were engaged in automation of the sports organizations for a long time – some of the first in the market. We have own specialized SmartSport system developed taking into account specifics of activity of sports clubs, and she, by the way, turns on also the analytical module for monitoring of performance indicators of activity of club and also for the analysis of effectiveness of commands, players and matches. Separately we work on development of the Crystal system which will be delivered both within SmartSport, and in the form of an independent product. This very interesting, focused on psychologists of clubs solution for conducting the most different psychological tests, the analysis of psychometric these players, assessment of compatibility of players in a command, researches of dynamics of change of characteristics of players in time, etc.

One more perspective direction for clubs – the analysis of Big Data from social networks. Here it is possible to analyze behavior of players, including potential, outside club, regarding observance by them of the mode and sports discipline or, for example, to try to control the fan movement, to timely apply measures in connection with some illegal actions.

TAdviser: Today development of the majority of IT platforms, whether it be ERP, CRM and or BI, is under the influence of such essential trend as social. So there are more and more elements of social networks or structural principles of social networks software gets into business. How does it find reflection in those products with which you work?

Olga Gorchinskaya: We already touched upon a subject of social networks. The most part of tasks in which inquire in the area of Big Data now are the analysis of social networks. It is necessary, for example, for banks for risks assessment, especially in the field of mass crediting, for more effective promotion of products and services, for increase in profitability of clients. At the same time simplicity of information extraction from social networks only seeming: not all data are open, special interfaces and gateways for collecting of these data from different information resources are required. For this reason we develop now cooperation with the companies which develop such interfaces and gateways decade. There is not a lot of them, we, for example, work with the companies "Analytical business solutions" and "Aykumen", having wide experience of implementation of similar solutions for security services of the different organizations, analytical services of the state and business companies. Such cooperation to us seems more effective, than attempts "invent the bicycle".

TAdviser: How do you estimate the level of penetration of data mining technologies on our market? Whether there is to them an interest from customers, some specific practices?

Olga Gorchinskaya: To data mining interest in the Russian market exists many years, but so far, as a rule, does not reach active implementations. Outside banking sector where data mining is used by risk management departments and the credit scoring, it is a little big promotions in this direction. A certain interest is in retail, but here, as a rule, it is about pilot implementations. There were experiments on use of Oracle Data mining in Toy Store department store for marketing tasks - such as implementation of target goods, increase in a customer loyalty. On sports analytics we also work in this direction.

Nevertheless, in the market there is a revival connected in many respects with the growing popularity of language of statistical investigations of R based on which Oracle created the product R Enterprise. We actively use both this product, and Oracle Data Mining in our projects. These means, for example, are used in the mentioned ROSEKO project where there is a lot of difficult analytics: clustering, self-organizing cards of Kokhenen and so on.

TAdviser: To what perspective BI and Big data directions will the FORS pay the main attention in the near future?

Olga Gorchinskaya: We continue to work in the field of profound analytics for the solution of practical tasks using statistical and mathematical techniques and also language R. We investigate possibilities of hands-off processing of texts for extraction from unstructured texts of the useful facts and their use in means of the class Data Discovery. So far there is not a lot of ready-made products in the field of linguistic processing in the market, besides, that in connection with specifics of Russian, we are guided only by domestic solutions. The interesting research direction represents the analysis of emotions (emotional intelligence) for the general data analysis about the person, for example, to assessment of his emotional relation to some subject or a subject without use of direct questions and questionnaires, only on the basis of special means of the analysis of physiognomics and a mimicry. There are specialized solutions which long remained niche, but in connection with development of big data in them big perspectives appeared, and we think over where and as they can be applied.

You See Also

Big Data - Directory of systems and projects