Provided is a process of identifying a name and boundary of a neighborhood based on web documents, the process including: extracting, via one or more processors, an n-gram appearing in a plurality of web documents; associating the n-gram with geographic locations associated with the web documents from which the n-gram was extracted; identifying a neighborhood by identifying a cluster of geographic locations associated with the n-gram; determining a boundary for the neighborhood from the distribution of geographical locations in the cluster; determining a name for the neighborhood from the n-gram; and adding the name and boundary of the neighborhood to a geographic information system.