Neural networks “under the hood”: the results of the hackathon of the “Honest Sign” system were summed up

The hackathon took place in several stages. Teams could choose which task to solve: develop services for efficient inventory, distribution, and sales management, or create a predictive market analytics system.

In the task dedicated to the development of a predictive analytics system, the WebBee team became the winner and owner of the main prize of 200 thousand rubles. Artem Kornilov, Denis Karabaza, Konstantin Barabanov, Roman Zubov and Igor Kuzmichev live in the Moscow region and work for the same IT company Webbee.

The team has created a market analysis service that allows you to make forecasts about future volumes of incoming and outgoing goods in circulation. Visually, the product is a web portal with graphs, charts and data tables, “under the hood” of which neural networks are engaged in forecasting. “All forecasts can be broken down by individual products, companies, regions and types of operations”– explained Igor Kuzmichev, captain of the team.

Government agencies can use such a system for planning revenue from the revenue side of the budget of the Russian Federation. But not only: the forecast of the availability of goods on the market will allow the antitrust authority to take preventive measures to avoid shortages of goods.

The consideration of the WebBee service surprised the jury: the team managed to develop a product that can not only predict the volume of trade, but also the size of the balance, which will allow the company to ensure the availability of products on the market in a timely manner. .

Second place was won by the Lucky Random team consisting of captain Egor Mulyukov, Alexander Evdokimov, Nikita Romanenko. The guys live in different cities in Russia, but this does not prevent them from solving complex problems together. Lucky Random has created a service that makes it possible to predict the demand for goods in regions based on historical data, as well as find deviations and anomalies in supply chains.

According to the participants, the emphasis in development was on creating a clear interface. They praised the quality of the data from the Chestny Znak labeling system: 68 PB of information about the goods, their movement and quality, useful for streamlining processes and making business decisions. Analysis of such a volume of information can help identify problems in the production, transportation and sale of goods, as well as improve the quality of goods and services.

The third-placed members of the Punk Butterfly team from St. Petersburg – Captain Alexander Shkolin, German Mityaya, Ilya Kupriyanov, Daniila Nemtsev and Alexei Poleshchenkov – proposed a solution aimed at combating monopolies in the market.

According to the team captain, the MARKING HACK hackathon was not easy for them in terms of coming up with an idea: “We discussed for a long time what benefits and how can be derived from the provided labeling data”. As a result, the participants came up with a product with extensive functionality, the characteristics of which are the use of economic metrics and the analysis of regional data.

In the task dedicated to the development of services for companies, the teams Viju, Leprechaun and Bumblebee took prizes.

Viju, who won first place and the main prize of 200 thousand rubles, offered an original solution to the problem – the service of an electronic sales representative. A team consisting of Stepan Ivanov, Andrey Solovyov, Bogdan Volik from the Moscow-based company “Information and Analytical System” devised and implemented a product for manufacturers who sell products through traditional retail.

“We analyzed the existing processes in traditional retail and found the use of data from the “Honest Sign” labeling system in terms of monitoring balances at points of sale”, – this is how the team captain Stepan Ivanov explained the emergence of the idea. According to him, the developed system optimizes the existing processes of the sales representative and thus saves money by increasing their productivity and efficiency. “The uniqueness of the solution lies in the application of automatic learning methods and classical regression analysis of the sales statistics of products by points of sale”Stepan Ivanov explained.

The development of this equipment has already been adopted: Svyatoslav Vilk, CEO of the Aqua Holding group of companies, during the hackathon briefing, suggested that the team test the “electronic sales representative” in “field” conditions.

The jury awarded second place to the Leprechaun team. Simple, understandable, accessible – the team’s decision left such an impression on those who watched the final. Using data from the Chestny Znak tagging system, they created a geographic demand analysis service for retailers to help increase sales of popular items in a particular city or area.

As the captain of the Leprechauns Alexander Andryushkin noted, the geoanalytics service is an inexpensive, but technologically advanced solution for small and medium-sized businesses. The idea for the service came to the team captain quickly, as he knew the data that the “Honest Sign” labeling system collects about products. He believes that this data opens up great opportunities for both IT specialists and businesses.

Third place was taken by the Bumblebee team from St. Petersburg. Participants created a product for retailers: a pricing and forecasting system that can be used to predict the demand for a certain product and select the optimal price for the product, at which revenue will be maximum.

According to team captain Yevgeny Pakhalyuk, the product solves procurement optimization problems and increases revenue by choosing the best price, and the system they created can make short-, medium-, and long-term forecasts using data from Chestny Znak and others. systems.