Westlake, Andrew (2003), Relational Databases and Statistical Processing Data Structures for HIV and AIDS Notification and Analysis, Computing Science and Statistics, 35, I2003Proceedings/WestlakeAndrew/WestlakeAndrew.presentation.pdf ,
Most statistical procedures are based on a single rectangular dataset, and much statistical data is conceived and stored in this form. Database professionals usually start from the relational database model for their data structure designs, effectively a set of related rectangular datasets. This allows the accurate representation of more complex structures, and, where this complexity is relevant, allows easier maintenance and manipulation of the data. This paper describes the restructuring of the processing and analysis system for the HIV and AIDS reporting system for England and Wales, and discusses the benefits that were achieved, in terms of easier processing and maintenance, more flexible manipulation, and faster reporting.