Determining the Popularity of Geolocation for ATM Placement

The aim of this graduation project is to create a web service for assessing the popularity of a geographical location for placing ATMs. The main task of the project is to develop an efficient and user-friendly service that allows users to quickly assess how suitable their chosen location is for placing an ATM. The web interface of the service is implemented as a Telegram bot. The principle of the Telegram bot's operation is as follows: the user enters the coordinates of the point where an ATM is presumed to be placed, and receives an assessment of the geographical location. This location assessment is presented as a popularity index, where a higher index indicates a better location for placing an ATM. The geographical location assessment is provided by a machine learning model based on the user's input. During the project, both traditional machine learning models and deep learning models were trained. Additionally, a comparison of models was conducted based on quality metrics, and the model with the best predictive ability was selected. To achieve a high-quality machine learning model, geo data was collected from open sources such as Open Street Map, Yandex Geocoder, and also socio-demographic data from websites that openly provide such information. This significantly improved the predictive ability of the model.