The developer created a desktop application called Krugosvetka to help a student memorize 82 geographic locations, including rivers, lakes, and straits. The software functions as a geography trainer that displays an unlabeled map and asks the user to pinpoint specific objects.

The application was built using Python and the Tkinter library, with Matplotlib used for rendering the map. It utilizes open-source Natural Earth vector data at a 1:10 million scale, allowing the program to operate offline without requiring an internet connection or API keys.

The system calculates accuracy based on distance using the haversine formula, allowing for specific tolerances for each object. For extended features like rivers, the program uses multiple anchor points to define the correct area, while providing hints and mnemonics after incorrect attempts.

The interface supports map navigation through click-and-drag functionality and zoom. Users can choose between a training mode, which offers hints and re-tests on missed items, and a test mode that tracks performance without providing assistance.

To ensure accessibility for other students, the developer used PyInstaller to package the application into a standalone Windows installer. All user statistics are stored locally in a JSON file within the system's local application data directory.