A method of transforming and canonicalizing semantically structured data includes obtaining data from a network of computers, applying text patterns to the obtained data and placing the data in a first data file, providing a second data file containing the obtained data in a uniform format, and generating interface specific sentences from the data in the second data file.