J. Delve & R. Healey
Data warehousing technology was developed in response to the perceived need to retain and re-structure commercial data collected from transaction processing applications, so they could be re-used and re-analysed for data mining and business intelligence purposes. Typical data warehouses now start with tens of gigabytes and grow into the terabyte range, although the technology can equally be applied to smaller data collections.
This paper aims to examine whether the potential range of uses of the technology for the storage, linkage and analysis of historical data is sufficiently large to justify the infrastructural overhead in establishing and maintaining a data warehouse. This is achieved by posing a number of questions, including the following:
Should data warehouses be project specific or can they encompass multiple projects?
In general, what are the advantages and disadvantages of a structured warehouse compared to a catalogued data archive?
Can data warehouses effectively handle multi-media data resources?
Can OLAP techniques yield useful insights into historical data?
Can warehousing and GIS methods be harnessed to build spatial data warehouses and if so, how could these be used?
These questions are examined in the context of an operational data warehouse (originally created as a research and teaching demonstrator), and other potential application areas. The demonstrator utilises research data on the 19th century anthracite industry of Pennsylvania.
|