The following presentations are licensed under a Creative Commons Attribution 4.0 International License.
Analysis of Sensory Data with R: Collaborative Open Online Course
These resources are part of a project that I've initiated with Margot Brard.
- Click here to get the Introduction resources.
- Click here to get the Quantitative approach resources.
- Click here to get the Affective approach resources.
- Click here to get the Qualitative approach resources.
Introduction to R
Please, look at these videos without moderation. It's important to know how to do things practically, but it's even more important to understand what you're doing. Please try to answer to the different questions related to the videos.
- Click here to learn how to download R from the CRAN website.
- Click here to learn how to install the Rcmdr package.
- Click here to learn how to load the Rcmdr package, in other words how to open it.
- Click here to learn how to save as a txt file with tabulation as separator.
- Click here to learn how to import a data set from Rcmdr.
- Click here to learn how to convert numerical variables into factors.
- Click here to learn how to get a boxplot according to a factor of interest.
- Click here to learn how to compare two means.
- Click here to learn how to select a subset of a data set according to a category of a factor.
- Click here to learn how to test a mean.
- Click here to learn how to import a data set from Rcmdr with "row names".
- Click here to learn how to install the SensoMineR GUI in Rcmdr (as well as the FactoMineR GUI).
- Click here to learn how to select a model in a multiple linear regression context, with the FactoMineR package (part 1).
- Click here to learn how to select a model in a multiple linear regression context, with the FactoMineR package (part 2).
- Click here to learn how to perform a PCA with the FactoMineR package.
Analyzing QDA data with SensoMineR
Please, look at these videos without moderation.
- Click here to learn how to assess the performance of your panel.
- Click here to learn how to assess the performance of your panelists.
- Click here to learn how to get a representation of your product space from the "raw" data obtained from QDA.
- Click here to download the slides on "Panel performance with SensoMineR".
- Click here to download the slides on "Analysing QDA data".
- Click here to download the "sensochoc" QDA data.
- Click here to learn how to describe automatically a qualitative variable.
- Click here to learn how to get clusters of individuals and how to describe them.
- Click here to download the orange data set.
Principal Components Analysis
The idea of the lecture is to bring the student to the notion of multivariate versus univariate or bivariate. In the first part of this lecture, the student is supposed to realize their first PCPA by hand, starting from a small dataset. The student will go through two very important notions, (1) centring the data, (2) standardising the data. Hopefully, she/he will understand the notion of distance amongst individuals (in our context multivariate distance).
- Click here to download the first part of the lecture.
- Click here to download the second part of the lecture.