PATN does pattern analysis. PATN is a comprehensive and extremely versatile, yet simple to use software package for extracting and displaying patterns in any type of complex multivariate data.
How does it work?
PATN generates estimates of association (resemblance, affinity, distance) between any set of objects described by a suite of variables (attributes). PATN then classifies the objects into groups, condenses the information into three dimensions and displays the patterns in your data graphically.
PATN also has a range of tools to help you to to identify the reasons for these patterns.
Where has it been applied?
'Pattern analysis' has been used in just about every area where data is available. To name a few disciplines - anthropology, botany, chemistry, ecology, fisheries management, genetics, hydrology, information technology, limnology, marketing, oceanography, psychology, socio-economics, taxonomy, writing analysis, zoology and more. If you have a set of objects described by some quantitative measurements, PATN can show you relationships.
A number of example analyses or case studies will added to over time to illustrate the utility of PATN for finding structure in complex multivariate data. Each worked example attempts to illustrate some of the common processes involved in pattern analysis. There is no single strategy that will ensure understanding of complex data. Analysis should be approached with an open mind (not easy), a willingness to question and pursue ideas and in some cases with some perseverance.
Two case studies are currently available with the associated (downloadable) data so that you can follow along-
- Medals (215KB). This example relates to the 'socio-economic' dataset that is included with distributions of PATN and is built into the open source version of PATN.
- Marine benthic community data (243KB). This worked example highlights the differences between polluted and 'control sites' in an Antarctic environment. The data is the result of a PhD project by Jonny Stark at the Australian Antarctic Division.