Data import: ConsumerCheck now also accepts data that are delimited by semicolon. A couple of the data tags were slightly renamed, to better describe the type of data: “Design variable” was renamed to “Product design”; “Sensory profiling/descriptive data” was renamed to “Descriptive analysis/sensory profiling”.
Prefmap: The X&Y correlation loadings plot , i.e. the preference map, has now an option for colouring according to how many consumers that are present in the different parts of the plot. The user can choose how many segments are to be used.
Prefmap & PLSR/PCR: The predicted values of Y are now available from calibration and validation. These data can be used to check how well the original data in Y are described by the model.
Conjoint: For structure 1 (struct 1) and structure 2 (struct 2) random effects and interactions containing random effects are no longer automatically removed when non-significant. Also, it is now possible to copy the conjoint Full model residuals and Double centred residuals to the Data sets tab by right-clicking. In this way the user can analyse the residuals further with for example PCA or PLSR/PCR.
GUI backend: ConsumerCheck 1.2.0 looks slightly different from 1.1.0. since we are using another backend for our graphical user interface (GUI). For those who want to know, we switched from wxPython to PyQT. The reason for this is that installation of ConsumerCheck for Mac and Linux users will be much more straightforward by taking out some obstacles that made easy installation really difficult. Currently, we are working on versions for Mac and Linux and should have those available very soon.
Text refinements: In various places we have improved the text in the GUI for clearer user instructions and error feedback.
To download the newest version please go to the Downloads section.