VISUALISATION / GRAPHICS
28 September 2011
Reviewed by Felix Grant
Software publishers vary in their approach to version numbering and frequency of issue. Some only issue major, full digit upgrades, and then at longish intervals. Others go for small but frequent issues. With some products, a third digit increment (as here between OriginPro 8.5 and 8.5.1) would signal a minor maintenance issue of interest to particular users; with OriginLab it indicates a useful general development, which will reward the effort of updating. So what has changed since release 8.5?
In Origin’s core function territory, data visualisation, there are a clutch of detail improvements across the board from using a dataset to specify axis ticks and box plot whiskers through colour palettes, context menus (which see useful revision across other areas as well), error par plots, histogram bin control and layers, to themes handling.
The growing analysis and statistics capability is expanded in several directions, some of them through the growing stable of 'gadgets' (of which more below). Two and three dimensional multiple peak fitting developments are part of an extension to the already versatile Levenberg-Marquardt based nonlinear fitting facilities. The existing Query Builder is augmented by a new SQL editor, allowing user controlled interrogation (including LabTalk variable substitution) of ActiveX data Object or Open Database Connectivity sources. The worksheet benefits from process supplements and usability tweaks, additional third-party file import formats and custom time format refinement.
As with other recent releases, the most intriguing feature remains the growth of Origin’s time saving and practice encapsulating exploratory productivity library of 'gadgets'. Each of these, essentially, is a way to identify visually a region of interest (RoI) within a visual display and them apply analytic tools to data which that RoI contains. The two new arrivals in 8.5.1 are Cluster and Quick Peaks, though two others (Statistics and Digitiser) have been developed. All of them do pretty much what it says on the tin.
The cluster gadget applies data brushing and management to visually identified clusters of points within a plot. Essentially, it is a way to isolate a coherent subset of data from one or more sets for separate examination or, alternatively, for exclusion from examination. Selected clusters can be copied, masked, cleared (by cell or by entire row), and other wise manipulated as an entity, with live display of basic statistical descriptors. Those with a facility for thinking in terms of arrays and factors (or 'cube heads' as my 15-year-old neighbour witheringly dubs us) will do this sort of thing with sorting and filtering. For most human beings, however, especially those who gravitate towards a visual exploratory tool for their environment for their analyses, the cluster gadget is much more intuitive and, therefore, much more productive.
The quick peaks gadget (peaks can be positive, negative – it can, in other words, be applied to troughs – or both) takes data within the RoI, identifies peaks (using a menu of five methods and a range of other options), and applies various methods to their study by mouse click. At the simplest level, the peaks can be expanded to make them more visible. Integration calculates area beneath them (either to the x-axis or to a local baseline for the RoI, which the gadget offers to determine for you). A number of other descriptors are also included: centroid, height, half-height widths, right or left half peak widths, markers, and so on. There are sundry other controls, too numerous to mention here, which together allow considerable fine tuning.
I habitually show new software packages to a wide range of people, to see how real users react to them, and few features have aroused so much enthusiasm across the spectrum from primary school to post graduate researchers than the gadgets. These two were no exception, triggering enhanced and focussed exploratory activity at every experience level.