Geographical Analysis of Journal Digital

This project performs geographical analysis of Swedish newsreels from the Journal Digital corpus (1930-1965) using Named Entity Recognition (NER), geolocation validation, and geocoding. The pipeline extracts location mentions from video subtitles, validates them against the GeoNames gazetteer, and converts them to geographic coordinates using Nominatim. Results include validated location datasets, temporal analysis of geographic mentions, regional distribution statistics, and JSON exports for spatial visualization. The project provides tools for preprocessing NER data, analyzing location patterns over time, and generating interactive visualizations of geographic coverage in Swedish newsreel production.