Geoprostorový doporučovací systém pro krátkodobé výlety na základě předpovědi počasí, kvality ovzduší a preferencí uživatele

Abstract

This work designs and implements a web-based geoinformation application recommending tourist locations within a 50 km radius of Ostrava based on weather forecasts, air quality data, and seasonal climate norms. The core of the system is a penalty-based scoring algorithm that evaluates each of the 363 locations with a score from 0 to 100 according to deviations of forecasted meteorological conditions from monthly historical averages; the scoring incorporates six factors and supports personalisation through userdefined preferences. The application is built on the Django framework with the GeoDjango extension, the PostgreSQL/PostGIS spatial database, and the Leaflet.js library; meteorological data are retrieved from the Open-Meteo API.