Abstract
This paper motivates, documents and evaluates the
process and results of converting the Norwegian
Petroleum Directorate's FactPages, a well-known and
diverse set of tabular data, but with little and
incomplete schema information, stepwise into other
representations where in each step more semantics is
added to the dataset. The different representations
we consider are a regular relational database, a
linked open data dataset, and an ontology. For each
conversion step we explain and discuss necessary
design choices which are due to the specific shape
of the dataset, but also those due to the
characteristics and idiosyncrasies of the
representation formats. We additionally evaluate the
output, performance and cost of querying the
different formats using questions provided by users
of the FactPages.
Users
Please
log in to take part in the discussion (add own reviews or comments).