ITMO: Shop Placement Platform

2024: Building a store revenue forecasting platform

Undergraduates ITMO , together with the company's experts X5 Tech , have created a platform that, based on geodata information trade and area, can predict the store's revenue in a certain location. The application helps to quickly analyze the location of a potential supermarket. The application was tested by managers and analysts of the retail network X5 Group in working with real commercial cases. The university announced this on September 20, 2024.

The work of the Shop Placement Platform application is based on algorithmic machine learning. This is software code that analyzes large amounts of data and finds patterns in them. For example, this technology is used in recommendation systems.

The algorithm was trained on real commercial data on operating stores of the Pyaterochka retail chain. The basis of the "scenarios" of forecasting how much a new store can "earn" was based on the dependencies between the location of the existing store, its characteristics and revenue indicators of existing stores.

The main part of the application is an interactive map with areas of profitability of stores. Areas painted red indicate the perceived low economic effect of the outlet in this space, green indicates high. The user selects a point on the map and adds characteristics of the retail space. For example, area, type of building (residential or office), whether alcohol and others will be sold. Further, based on this data, the algorithm generates a "report" on the potential economic effect of the store and makes recommendations regarding construction here. By comparing the data on several points, the user determines the optimal location of the new store. The platform cannot replace an analyst or manager, but it can become a tool for more efficient and faster work, "said Daniil Solovyov, one of the developers of the application, a master's student in the joint educational program on artificial intelligence ITMO and Napoleon IT AI Talent Hub.

To confirm the accuracy of the analysis results, the developers compared the platform data on existing retail outlets with their real cost-effectiveness reports. In most cases, the indicators were close.

For six months, students, under the guidance of experienced experts, created digital products. As a result of the internship, the best undergraduates joined the company's team.