[EN] The accessibility of official statistics to non-expert users could be aided by employing natural language processing and deep learning models to dataset lexicons. Specifically, the semantic structure of FIPS codes ...