Nowadays the internet is overcrowded with huge amount of textual data – various news portals, blogs, forums, social networks, encyclopedias, books etc. New methods are required for information structuring and searching, linking additional information to context thus making it easier to find and perceive the information, as well as making it more useful. One of the contexts that can help to solve this problem is geographical context. This work is devoted to the research of automated textual data geocoding (or geoparsing) matter and solutions thereof. Based on the scientific research methodology a system prototype is developed that is capable of analyzing texts in automated mode, recognizing geographic information in those and marking geographical names. The research results and the system prototype can be used as an automated approach to a text grouping by its geographic context (for example, in news portals); as well as it can be used for aggregation purposes. Such system can automatically link the geographic information (coordinates, addresses, zip codes, cities, states, even geocoded images and videos, etc.) to appropriate words in any text containing geographical names. All this can allow a user both getting relevant additional geographical information about the text, as well as, performing searches in the text using the abovementioned geographical data.