Dialectology is a branch of linguistics that analyzes the geographical and sociolinguistic variation of languages in space. However, traditionally the spatial component in dialectology is usually limited to the location of data collection without considering other geographical aspects that can be measured.
According to this article, it shows the results of the research to determine the existence of a relevant quantitative relationship between linguistic variants and geographic variables, through the design of a spatial model that incorporates lexical data from the Linguistic-Ethnographic Atlas of Colombia (Atlas Lingüístico-Etnográfico de Colombia (ALEC)), metadata and information of different geographical phenomena. Among the results, the evaluation explained by the spatial autocorrelation index suggested that the variables: agroclimatic suitability, precipitation, geographical distance and access roads show the highest bivariate spatial dependence in relation to the linguistic distance calculated with the Relative Identity Index (IRI). The treatment of this interaction within the mixed geographic autoregressive spatial regression model confirmed this dependence, thus corroborating the relationship between language and geography.