Semantic Type Analysis

The goal of this project is to assign semantically meaningful types to collections of data values. While types such as strings provide some information, a more useful type for analysts provides additional semantic information. For example, data values such as Kampala, Lima, and Shanghai may be labelled with the semantic type city. This is not a completely new problem, but we aim to consider how the problem of semantic type analysis relates to semistructured data in order to exploit the relationships between nested data values.