This document presents a system that aims to derive location information from visual content in videos without geo-tags. It divides the world into regions based on different criteria like temperature and biomes and uses visual similarity measures to match videos to these regions. Initial tests on 500 videos using a 22 biome classification achieved a 12.17% accuracy, better than random chance of 4.55%. Future work will focus on only using outdoor videos and excluding indoor images which provide noisy information.