In the age of rapid technological development, more and more people prefer to receive goods and services online. Marketplaces and online stores are looking for optimal and convenient ways to deliver their services to customers.

My Postamat helps e-com players develop new ways to provide services with maximum proximity to customers, using the model of sharing a network of parcel terminals.

For these purposes, a service was required to search for optimal and popular locations for the placement and relocation of a network of parcel terminals with the ability to generate a data model based on machine learning in real time.

The Yourfavorite.agency has developed a PostNet service that helps solve this problem as part of the My Postamat project. PostNet is an innovative project that uses data on localities, population density and demographic characteristics to determine optimal locations for parcel terminals.

Goals

Develop a geoservice that, based on many factors, determines the optimal locations for parcel lockers, allowing you to work with a large amount of data

Ensure maximum efficiency in finding optimal locations for the parcel terminal network through the use of machine learning and real-time data updating.

Provide users with a simple and intuitive interface that allows them to select optimal data using various filters.
Results

We developed the My Postamat geoservice, which made it possible to accurately analyze many factors, including population density, traffic, and other parameters for the optimal placement of parcel terminals. We implemented a convenient interface for analyzing the effectiveness of the parcel terminal network with the ability to independently search for optimal locations.

Analytics

The project team carried out analytics, identifying patterns that determine the demand for places for parcel terminals. These patterns served as the basis for an accurate model capable of predicting optimal locations for parcel terminals. Various data was used, including information on transport, educational institutions, shops, and other factors collected from official sources and from parcel terminals in real time. A distinctive feature of analytics in My Postamat is the use of machine learning, which is updated based on current data from parcel terminals. This ensures high accuracy of the model even with changes in urban infrastructure. The result is an effective tool for determining the optimal locations for parcel lockers, supported by up-to-date data and high accuracy thanks to machine learning.

Design

The map is the central element of the MyPost system interface. It allows users to quickly and easily find the desired parcel locker. To ensure ease of use of the map, Martin technology is used, which allows data compression and the use of lightweight protobuf files. This significantly reduces the load on the network. To visualize the data on the map, a heat map of a contrasting shade with a white background is used. The heat map allows users to quickly and easily see where parcel terminals are located. The card also allows users to:

• Search for parcel terminals by address or name.

• Filter parcel terminals according to various criteria.

• View information about the parcel terminal.
  In the future, the map may be supplemented with information about how busy the parcel terminals are and integrated with other services.

Development

Client Part (Frontend):

• React: Creating a flexible and scalable user interface.

• Ant Design: Using UI-kit for beautiful and functional interfaces.

• Tailwind CSS: Accelerated styling according to atomic design principles.

• Zustand: Centralized state management for cleaner, more maintainable code.

• Maplibre + react-map-gl: Map display with support for geometric operations.

Backend:

• PostgreSQL + PostGIS: Storage and processing of spatial data.

• Vector tiles: An efficient way to visualize spatial data.

• Django: Python framework for business logic.

• nginx: Processing user requests.

• Docker containerization technologies:

  • Microservices: Functioning of components as isolated containers.

Result: Effective interaction between the client and server parts in the innovative geoservice My Postamat.

Architecture

The My Postamat system is a client-server architecture using microservices. The client part of the system interacts with the backend through HTTP requests. The backend consists of several microservices, each of which is responsible for a specific function. The system is developed using modern technologies that provide flexibility, scalability, and efficiency of the application.

Finally

Now the service is working and fulfilling all the ideas that were put into it. The My Postamat system is a successful example of using modern technologies and data analysis methods to solve a complex problem. The system can be used as a basis for the development of similar systems in other cities and countries.