Good evening! Today, the first talk about data analysis in our blog. For many, this is a very hot topic. However, there are not so many really useful analytics meetings and conferences in Belarus. On October 25, 2014, the first DataTalks took place in Minsk. DataTalks are informal meetings of data analysis specialists. For participants, this is a great opportunity to learn about the experience of analytics in companies operating in the global and local markets, as well as to get acquainted with experts in the field of data analysis from various industries.
Speakers from Wargaming, Yandex, Dmlabs.org and Nanotech, with examples of the problems they solve, explained the general patterns and applicability of algorithms that are important to know and use when analyzing data in any industry. Questions to the speakers turned into professional discussions, which could not stop even the beginning of the next report.
To tell more about how the idea of creating DataTalks came about, we asked Sergey Kadomsky , the ideological inspirer of the analyst community and the director of Wargaming research:
The idea to organize such an event appeared after we began to look for experts in the field of data analysis in Belarus to exchange experience. It turned out that there are few such people and many of them do not even realize that each other exists. At the same time, the experience of our company indicates that data analysis is not just a competitive advantage, but a prerequisite when creating products and doing business in the international arena. I am sure that our events will have a beneficial effect on the professional development of data analysts, and, as a result, on the growth of the performance of Belarusian companies both in the domestic market and on the world stage. I am very pleased that with the support of Wargaming, we were able to bring together 200 specialists who are interested in and professionally engaged in data analysis. On behalf of the company, I want to thank the speakers and all the participants of the event. We are planning to hold the next meeting in February 2015, we will be glad to see you again!
Let's move on to speeches. The first word on DataTalks took Mikhail Levin. ')
Mikhail Levin, Head of Big Data Analysis Group, Yandex Data Factory "How to make millions from your data and machine learning." Mikhail spent a little excursion into the world of Yandex.Direct. This is the system of placing relevant contextual advertising in each search query, thanks to which Yandex gets the main profit. He spoke about the basic principles of the operation of Yandex.Direct and how to implement Machine Learning not to be mistaken, testing the effectiveness of new algorithms on real users. You can view and download the presentation of Michael here .
Ksenia Petrova, COO dmlabs.org “Data mining in practice.Pitfalls of data analysis " From the point of view of her experience, Ksenia told about common mistakes that analysts face. It was about how to properly build processes to avoid these mistakes, and thereby save time and money. To do this, researchers should strictly monitor the quality of the data, study the experience of other analysts, endlessly improve the models and most importantly, remember that the business problem should always come first. You can view and download the presentation of Xenia here .
Sergey Kadomsky, Wargaming Research Director "How to sell analytics to business" Sergey told about the approaches with which the analyst can be built in the business as a whole, and how this happens specifically in Wargaming. It was about the features and principles that should be remembered by analysts in their work and communication with the business in order to get a result that suits both parties. You can watch and download Sergey’s presentation here .
Ivan Myakishev and Alexey Yurkevich, data scientist of Wargaming "Compromises in Data Science" Every specialist is faced with this in his professional activity: at some point we understand that the theoretical knowledge obtained in the university and from special literature does not work. Real life is not an example of an online Machine Learning course. Ivan and Alexey talked about the main trade-offs that business and analysts have to work together. These are trade-offs for speed, accuracy and simplicity. The importance of these compromises was shown by examples from their experience in Wargaming. View and download the presentation of Ivan and Alexei can be here .
Alexander Grinchuk, Associate Professor, IBMT BSU, Nanotech LLC "The practical application of Data Mining technologies on the example of real business problems in various industries." Alexander spoke about the national characteristics of the Belarusian data analysis market. Using the example of real business problems, he showed the main problems faced by experts when implementing Data Mining in Belarus. This is a lack of a clear understanding of the need for processing, storing and analyzing data from a business, and a low level of processing quality, often offered by business experts. You can watch and download Alexander’s presentation here .
Round table on current issues of data analysis. At the end, a round table was organized for the most active and interested specialists. The main topic of discussion was the prospects for the use of Data science in the business of Belarus and Russia.
We hope you are interested in DataTalks. Join the analyst community on LinkedIn , where you can discuss interesting professional cases and useful materials. Would you like to discuss a specific topic in the next DataTalks? Or would you like to submit a report? Email us at meetup@wargaming.net