Zpracování prostorových dat

Abstract

This bachelor thesis presents the design and development of a software solution for spatial data analysis, whose primary purpose is to find optimal routes within a street network of a selected geographic area. The work is based on real-world geospatial datasets in Czech formats — a street network in GeoMedia GML format using the S-JTSK coordinate system, and points of interest exported from OpenStreetMap in KML format. The result is a web application consisting of a REST API backend implemented in .NET 8 (ASP.NET Core) and a frontend interface built on the Nuxt 4 framework. The backend handles loading and in-memory processing of spatial data, coordinate projection, construction of a graph representation of the street network, and shortest-path finding using Dijkstra's algorithm. The system also supports filtering of points of interest within a given geographic bounding box and visualization of the route neighbourhood using BFS (breadth-first search) over a street adjacency graph. The frontend renders results on an interactive map using the Leaflet.js library. A key output of the system is a Prolog export: streets, intersections, and segments are represented as facts; for each computed route the system automatically generates inference rules enabling queries over the route (first/last street, turn detection, proximity to points of interest) and a narrative_step/2 predicate that derives a natural-language description of each route step. This thesis lays the groundwork for a future system capable of identifying a route from a plain-text description.