A number of specialist systems for data entry have appeared over the years to complement the analysis facilities available in the major statistical packages. The arrival of high quality relational database systems for PCs, particularly Microsoft Access, means that good general purpose data management tools are now available in the Windows environment. Taken together with improved tabulation and graphics facilities in several spreadsheets, there are now many choices for constructing flexible, high quality systems for data entry, checking and presentation, complete with explicit meta-data and with full control and documentation of data quality, if required. The Internet offers new options for distribution of or access to results and ad hoc statistical information.
Statistical analysis is generally well served by standard packages, with excellent new versions of the established products (such as SAS, SPSS, Genstat). Competing with these are newer arrivals such as Systat (now owned by SPSS), WinStat, StatLab and Unistat which aim for easier use for basic tasks. Systems with full statistical programming languages like S-Plus, Stata and XLisp-Stat can provide more flexibility for implementing graphics and new statistical methods.
However, no package offers full flexibility for complex tabulation and for the handling of complex data, nor the special facilities needed for large-scale processing systems as found in government census or statistical offices. For these more complex situations a combination of relational database and statistical systems is often appropriate.
For the dissemination of summary data things may be at a change point. The 'Pivot Table' feature in MS Excel (from 2000) is much improved, and can be linked to summary tables in SQL Server 7 (and other sources). This provides a good model for the dynamic display features needed for browsing or exploration of summary data. Two smaller companies, Ivation with their Beyond 20/20 product, and Space-time Research with SuperTable, have introduced new software that brings a more statistical approach to the same area. Both also support the processing of large-scale raw data to produce the summaries.
Page last updated 09 May 2004.