The BIG DATA 2018 Business Forum is
the year's primary event focused on
big data and analytics

A significant, if not the largest share of big data is telemetry, which is data streams being captured by various sensors. Telemetry is special in that normally there is no need to keep all of the data as it either could be processed on-the-fly or pre-processed deeply enough to extract data (or events described by data) worth to be retained for further analysis. Sometimes outcomes of real-time pre-processing can offer critical insights requiring immediate response, e.g. by indicating emergency state in a cyber physical system.

Tasks like that represent a significant challenge for modern computing systems as majority of architectures are designed both at hardware and software level for fast real-time data processing, while the density of data streams in state-of-the-art cyber physical systems is constantly increasing as issues addressed by them grow in complexity. Initiatives in the area are already being implemented both by research community (with one example being the collaboration between Russia's Higher School of Economics and CERN) and by the industry. What approaches are used in this area currently, and what direction can the development of hardware and software for big data streams processing take under an edge computing model?

The explosive growth of robotics, artificial intelligence systems and big data technologies combined with globalization create an environment where all products and services compete on a planetary scale. As global platforms are being established with the digital economy paradigm taking over, only truly innovative products and services can be profitable, creating jobs and generating income for the government.

With technology being the main pillar of growth, the ability to rapidly innovate will be a key competitive advantage of people, businesses, and governments. Thus, «exponential technologies», i.e. the ones that enable exponential growth, will be critical to success of companies and even entire nations. Such technologies, however, could be applied only to existing assets. The potential of digital economy in Russia would be unlocked not by accelerators (considering that in the current environment, investment portfolio management is a monumental challenge), but rather by «expolerators» (exponential accelerators), capable of combining the resources of state-owned corporations with the agility of startups, financed through the monetization of intellectual property and its further global commercialization. This will be facilitated by the technology potential and human capital available in Russia.

The presentation describes practical approaches to Big Data, Advanced Analytics, and Machine Learning application to production problem solving. Case studies include solutions developed for NLMK steelmaking efficiency improvement. The techniques and tools covered in the presentation can be used for all sorts of production. Peculiarities of the mentioned technology application in complex manufacturing vs. banking and Internet segments will also be explained.

SIBUR will share their expertise in applying Big Data tools for petrochemical production problem solving; supporting Industry 4.0 approaches with Big Data and IoT; and technical and organizational details of Big Data utilization in digital production.

IT industry progress drives significant improvements in education quality. Discussion on benefits, drawbacks, and transformation opportunities of traditional classroom-based education has been held for a long time now. Having used mobile devices «from the cradle», modern-day students can often find more complete, accurate, and up-to-date information on most theoretical and practical subjects than their lecturer can offer. Moreover, sometimes online information is delivered in a more engaging and efficient form than in classes. Are students interested? Are they able to follow their teacher? Is the teaching material being delivered in an understandable way? How much the students are involved in the educational process in classes? These questions come to the forefront in the era of digital education. Until recently, there was virtually no ability to monitor student engagement level, considering that, for example, only in Moscow buildings of the Financial University classes are taught daily from 8:30 am till 10:00 pm in over 500 classrooms.

The Financial University will outline the development and deployment of a cloud service that continuously analyzes feeds from video cameras installed in classrooms, identifying students" faces with machine learning models, detecting emotions, and determining engagement level, followed by aggregation of the data by student groups, departments, courses, etc., and visualization through interactive dashboards.

LED revolution has reinvented the lighting industry. Lamps turned into electronic equipment—and got smart faster than many other elements of urban/industrial infrastructure, joining the IoT and Digital technology revolution.

Data, its collection, management and analysis all play a key role in new business models. Is device status data enough for Lifecycle Services provision? Can the volume data aggregated throughout the lighting system lifecycle have a business value of its own? Heatmaps and the facility visitor behavior analysis are merely two of the most obvious examples.

What do, then, manufacturers base their future success on: should they be selling devices, services, or data?

I will speak about how the use of big data changes drug development landscape and administration practices. Big data analytics is already being used for drug discovery and development, as well as for evaluating medication efficiency, safety, and economic feasibility. I will highlight issues that big data help address, as well as limitations of big data use.

Using case studies, Natalya will speak about employing big data analytics in HR, as well as about the ensuing changes and new HR capabilities in organizations. Even today, we can see how exponential technology growth is transforming HR responsibilities, sometimes changing the role into completely unexpected shapes.

The revenue forecasting solution was created for Finnish-Russian apparel retailer Finn Flare by IT Pro based on Microsoft technologies and Azure cloud platform using pre-processed historical internal and external data. The solution has enabled the increase of frequency and performance of creating a short-term sales forecasts facilitating faster managerial decision making.

The keynote will specifically highlight:

• The business value of the implemented solution
• The matters of quality and sufficiency of the data used for the forecasting
• The process of identifying internal and external factors impacting the forecasted variables
• The use of cloud-based technologies for building the forecasting models

The revenue forecasting solution was created for Finnish-Russian apparel retailer Finn Flare by IT Pro based on Microsoft technologies and Azure cloud platform using pre-processed historical internal and external data. The solution has enabled the increase of frequency and performance of creating a short-term sales forecasts facilitating faster managerial decision making.

The keynote will specifically highlight:

• The business value of the implemented solution
• The matters of quality and sufficiency of the data used for the forecasting
• The process of identifying internal and external factors impacting the forecasted variables
• The use of cloud-based technologies for building the forecasting models

A major challenge faced by almost every company is not the lack of tools, but rather shortage of professionals capable of doing something valuable for the business. Almost all currently available tools are designed for programmers, which are scarce today and would be in short supply for years to come. The Loginom platform makes advanced analytics accessible to mainstream users, but primarily to those who understand business.

Modern manufacturing businesses increasingly consider creating digital twins of their factories, yet with such an approach requiring complex mathematical modeling and expansive knowledge of various system's levels, this leads to the project being too resource costly. Machine learning and big data analytics technologies, meanwhile, in many ways seem to be a silver bullet for addressing data-intensive tasks without requiring deep knowledge of the process. Visiology will speak about using both of the approaches to enable real-time optimization of selectivity for chemical process manufacturing as part of a project implemented for SIBUR.

Stop being afraid of big data! A major problem with big data in real-world tasks is too little usable data fit for analysis, and too much barriers preventing access to such data.
"Why haven't you yet began accumulating data?"
Let's talk about how to begin collecting your data today so that tomorrow you'd be able to use it efficiently.
What data is critical to your business?
Let's talk about data inertia, because if we fail to collect data, we lose money.
In his keynote speech, the Naumen expert will address the issue of managing data flows of an enterprise. He will break myths about the nature of big data, and, in proposing an approach to making data work, he will discuss architecture of a solution for orchestrating data workflows using actual projects as examples.
Finally, he will outline a case study and cost advantages of using distributed computing for big data based machine learning and semantic analysis technologies.

Being on a steady upturn, Russia's agricultural sector is currently one of the economy's major growth drivers. The government of Russia believes in growth via digital transformation of the agricultural sector and use of advanced technologies such as precision farming, robotics, IoT, agricultural scouting (continuous crop improvement through analytics), and others. Things get even more complex in cattle farming, where the variety of solutions being used is further impacted by biotechnology.

All the new solutions mentioned above rely on heterogeneous streams of big data, requiring modern decision support and analytics technologies based on machine learning, neural networks, etc. Still, creating ontologies as a crucial tool of comprehensive big data analytics, has for now lacked attention and is not being automated.

Ontologies are obviously needed in public sector as well, such as the UNSPSC ontology, which could be suitable for use by various government agencies in Russia.

The keynote offers a breakdown of big data ontologies creation process for agricultural sector, laying the groundwork for bringing ontology-related projects from experimental stage into applied use area.

Artificial intelligence technologies will enable automated solving of some complex problems, which previously could be efficiently tackled only by humans with a certain level of mental abilities. The scope of the problems include image recognition, making decisions in unforeseen situations, knowledge discovery, and some others. However, a full-scale replacement of a human with an autonomous system is feasible only provided there is a way to verify that the system's functionality matches the abilities of a highly-qualified human operator when addressing a specific task.

Many businesses in Russia have deployed big data management platforms. Still, with the technology being too complex for the majority of business users, far from all Big Data projects have achieved success. The keynote will highlight approaches and best practices helping to significantly ease access to and management of big data for company employees having no advanced Data Science or Machine Learning skills.

Modern society often poses a question, which phase of the 6th wave of innovation are we in, and what is the relation between the innovation waves theory and digital economy? In my opinion, the answer is that we are already on the verge of the growth phase, for which all the prerequisites are in place, including the large amount of accumulated information and modern big data technologies, although our mindset is not ready yet for making decisive moves.
Digital revolution is already ongoing, and it brings forward a number of messages determining its further way, with us being direct participants of the transformation. In our case, it is the Digital Economy state program of Russia that sets forth 8 directions for rapid progress. In my presentation, I will outline the way our company makes its practical contribution to the economy digitalization.
While the explosive growth of big data market created a need for advanced analytics in a multitude of organizations, the understanding of objectives and approaches to achieving them is lagging behind the urge to keep up with the latest trends. "Let's load the data now, and understand what to do with it later," – that's a typical approach in many companies beginning to use the technology, with that among other things being the reason for many project failures. In his keynote, Evgeny Stepanov will describe the vision and strategy of Micro Focus for building modern future-oriented data warehouses and analytics systems.

Shipping expenses is one of the price components of all goods. By lowering shipping costs, businesses maintain higher profitability and gain competitive edge. Fast Data technologies help to achieve those objectives more quickly and efficiently.

The keynote will focus on an IT solution deployed in a major cargo shipment company, designed to improve cost planning, transportation forecasting accuracy, and real-time warehouse stock analysis. The solution processes over 500 transactions per second, enabling real-time week-ahead stock capacity forecasting for 180 warehouses all over Russia, down to specific cargo type.

The solution's architecture is based on Hortonworks Data Platform along with Apache Spark, Hadoop, Kafka, and the SnappyData database.

The keynote will outline architecture and technical details of Arenadata Unified Data Platform, a solution that allows querying of data at various technological layers (In-memory, MPP, Hadoop), using a single entry point.

Performance of specific queries, query plans, and response times will be discussed.

The keynote will outline architecture and technical details of Arenadata Unified Data Platform, a solution that allows querying of data at various technological layers (In-memory, MPP, Hadoop), using a single entry point.

Performance of specific queries, query plans, and response times will be discussed.

Big data is a rather valuable turnover subject that should be regulated by civil law in the era of digital economy. As of now, however, the big data turnover is not regulated in any explicit way, so the issue of big data ownership stays open-ended. Some may think that big data has no owners, so it could be used freely. Others believe big data is owned by technology companies who spend lots of money collecting it. Human rights advocates, meanwhile, think only users themselves can be the owners of their data. There is also a view that big data should be owned by the government. There is, however, no universal approach as of now.

Perhaps, the issues indicated above should be addressed not by legislation, but rather through legal precedents, with court rulings capable of being more flexible in adjusting to new challenges. For this very reason, we can regard as a truly landmark the ruling of Moscow Arbitration Court, effective from January 29, 2018, in VKontakte vs. Double Data, who developed software to analyze large VK datasets for commercial purposes, and subsequently sold information to banks, who in their turn used the data for establishing credit ratings.

The keynote speaker will discuss the precedent-setting case in detail, as well as outline alternative approaches (including GDPR-based) to estimating which personal information should be deemed as subject to legislative protection, and how the two types of digital data interrelate.

With data volumes and formats variety growing, the industry is offering many solutions for handling them. That, in turn, has made it more difficult to select right solutions for various tasks in a way to ensure interoperability and performance of the solutions throughout their entire lifecycle.
As a response to those challenges, we have introduced our Arenadata Unified Data Platform concept. Its developers were tasked with meeting straightforward requirements:

• support for any data type/format,
• support for ingesting data in the required mode,
• being capable to scale along with the growth of stored data without requiring system architecture upgrades,
• modular structure,
• being capable of interoperating with any tools and analytic workload types.

Advanced analytics methods can revolutionize business operations. The keynote discusses the significance of creating a big data storage system, data curation, the necessary software, and analysis methods. All those aspects are essential for moving from traditional business intelligence to creating analytics data lakes.

The BigData 2018 conference will be the event where Arenadata Unified Data Platform will be introduced for the first time ever.