basicR4FOOD

• Program
• Installation of R and Rstudio
• Datasets
• Additional material
  • Notes from the Data Analysis in Food Science course
  • Cheatsheets
  • Videos
• Import data
• Summary stats
• Simple plots with ggplot2

Course materiale for intensive R course at Dept FOOD UPCH

Program

Where: TBA

When: Tuesday, October 27th, 2020

10-11 : R environment, Import data, data.frame as the golden object, simple plot (hist())

11-12 : Descriptives (table(), aggregate())

12-13 : Lunch and Data-social

13-14 : ggplot2 [scatterplots, barcharts, boxplot, jitter]

14-15 : linear models and t.test

15-16 : PCA with ggbiplot and extraction of scores/loads -> into data.frame and ggplot2 plots

Installation of R and Rstudio

Take a look in the file installation_guide

Datasets

Datasets for this course is included the dataset folder /data

• Coffee temp (Temperature_Coffee.xlsx): A sensorical panel accessing the sensory properties of coffee brewed at different temperatures.

• GC-Cheese (cheese_aromas.xlsx): Aromaprofiling of chesses produces under different conditions over different storage times.

• Mouse (Mouse_diet_intervention.xlsx): Dietary effects of whey, casein as well as high low fat diet on growth and biomarkers in a mice model.

Additional material

Notes from the Data Analysis in Food Science course

There is a lot of online available material on statistics using R. The Notes for Data Analysis in Food Science covers basic data analysis in eight weeks. Data for most of the content there is to be found here and can be downloaded directly in R using the devtools package:

devtools::install_github('mortenarendt/DataAnalysisinFoodScience')
library(DAinFoodScience)
data("maramaox") 

Cheatsheets

RStudio has made available Cheatsheets covering the utility of different packages, such as ggplot2. These are very condensed, but also very usefull for checking out what you can do with a particular package.

Videos

For the Data Analysis in Food Science course, there is a bunch of videos available going through theory and How to in R. These can be found on youtube here.

Import data

Download the excel datafiles to an appropriate location on your computer. Further, install the rio package (This you only need to do once).

install.packages('rio')

Make a script containing the stuff below, save the file in the same place as the data, and set the working directory to the source file location using Session > Set Working Directory.

library(rio)
X <- import('data/Temperature_Coffee.xlsx')

This dataset contains a total of 192 observations. That is data from a sensory panel of 8 assessors (Assessor variabel in the dataframe X), across 6 products - coffee at six different brewing temperatures (Sample in the dataframe X), in 4 replicates (Replicate in X), assessed on 8 attributes: Intensity, Sour, Bitter, Sweet, Tobacco, Roasted, Nutty and Chocolate represented as individual columns in X. These attributes are all scored on a 0 to 15 scale.

Summary stats

Lets calculate the mean for the attribute Intensity (one of the variables in the dataset) for each temperature setting

aggregate(X$Intensity, list(X$Sample),mean)

##   Group.1         x
## 1     31C  8.306250
## 2     37C  8.620313
## 3     44C  8.517188
## 4     50C 10.528125
## 5     56C  9.946875
## 6     62C 10.110938

... Now lets do it for all attributes.

aggregate(X, list(X$Sample),mean)

##   Group.1 Sample Assessor Replicate Intensity     Sour    Bitter    Sweet
## 1     31C     NA      4.5       2.5  8.306250 8.451562  9.487500 3.285938
## 2     37C     NA      4.5       2.5  8.620313 9.075000  8.873437 4.143750
## 3     44C     NA      4.5       2.5  8.517188 7.087500  9.281250 4.659375
## 4     50C     NA      4.5       2.5 10.528125 7.589062 10.115625 3.046875
## 5     56C     NA      4.5       2.5  9.946875 8.268750 10.598437 2.868750
## 6     62C     NA      4.5       2.5 10.110938 7.790625  9.632812 3.314062
##     Tobacco  Roasted    Nutty Chocolate
## 1 10.420312 7.884375 3.829687  8.057813
## 2  9.979688 8.695312 3.914062  7.650000
## 3  8.878125 8.882812 4.410938  7.856250
## 4  9.529687 9.660937 4.743750  7.659375
## 5 10.120313 9.693750 3.829687  6.820312
## 6  8.835938 9.787500 4.195312  6.510937

... And lets get some more summary stats returned as well

aggregate(X, list(X$Sample),function(x) c(length(x), mean(x), sd(x)))

##   Group.1 Sample.1 Sample.2 Sample.3 Assessor.1 Assessor.2 Assessor.3
## 1     31C       32       NA       NA  32.000000   4.500000   2.327951
## 2     37C       32       NA       NA  32.000000   4.500000   2.327951
## 3     44C       32       NA       NA  32.000000   4.500000   2.327951
## 4     50C       32       NA       NA  32.000000   4.500000   2.327951
## 5     56C       32       NA       NA  32.000000   4.500000   2.327951
## 6     62C       32       NA       NA  32.000000   4.500000   2.327951
##   Replicate.1 Replicate.2 Replicate.3 Intensity.1 Intensity.2 Intensity.3
## 1   32.000000    2.500000    1.135924   32.000000    8.306250    2.749479
## 2   32.000000    2.500000    1.135924   32.000000    8.620313    2.452052
## 3   32.000000    2.500000    1.135924   32.000000    8.517188    2.780450
## 4   32.000000    2.500000    1.135924   32.000000   10.528125    2.399225
## 5   32.000000    2.500000    1.135924   32.000000    9.946875    2.325800
## 6   32.000000    2.500000    1.135924   32.000000   10.110938    2.625549
##      Sour.1    Sour.2    Sour.3  Bitter.1  Bitter.2  Bitter.3   Sweet.1
## 1 32.000000  8.451562  2.699357 32.000000  9.487500  2.569706 32.000000
## 2 32.000000  9.075000  2.476502 32.000000  8.873437  2.559832 32.000000
## 3 32.000000  7.087500  2.268792 32.000000  9.281250  2.575560 32.000000
## 4 32.000000  7.589062  2.639059 32.000000 10.115625  2.401190 32.000000
## 5 32.000000  8.268750  2.850064 32.000000 10.598437  2.495075 32.000000
## 6 32.000000  7.790625  3.185842 32.000000  9.632812  3.132494 32.000000
##     Sweet.2   Sweet.3 Tobacco.1 Tobacco.2 Tobacco.3 Roasted.1 Roasted.2
## 1  3.285938  1.758126 32.000000 10.420312  2.430945 32.000000  7.884375
## 2  4.143750  2.022086 32.000000  9.979688  2.092030 32.000000  8.695312
## 3  4.659375  2.759542 32.000000  8.878125  2.833638 32.000000  8.882812
## 4  3.046875  1.528015 32.000000  9.529687  2.791198 32.000000  9.660937
## 5  2.868750  1.672392 32.000000 10.120313  2.494603 32.000000  9.693750
## 6  3.314062  2.308233 32.000000  8.835938  3.016253 32.000000  9.787500
##   Roasted.3   Nutty.1   Nutty.2   Nutty.3 Chocolate.1 Chocolate.2
## 1  2.735502 32.000000  3.829687  1.885080   32.000000    8.057813
## 2  2.543155 32.000000  3.914062  1.643326   32.000000    7.650000
## 3  2.125331 32.000000  4.410938  2.158245   32.000000    7.856250
## 4  2.256558 32.000000  4.743750  1.817955   32.000000    7.659375
## 5  2.763960 32.000000  3.829687  1.555109   32.000000    6.820312
## 6  2.664552 32.000000  4.195312  1.609917   32.000000    6.510937
##   Chocolate.3
## 1    2.202971
## 2    2.518928
## 3    3.709572
## 4    2.980324
## 5    2.767696
## 6    3.090770

Simple plots with ggplot2

First, import the package ggplot2

install.packages('ggplot2')

Now let's make a plot of the Intensity (on the y-axis) versus the temperature (on the x-axis) and color all the responses according to the Assessor

library(ggplot2)
ggplot(data = X, aes(Sample, Intensity, color = factor(Assessor))) + 
  geom_point()

Now, lets join the points within each assessor (and replicate), and further, plot each replicate in its own panel.

ggplot(data = X, aes(Sample, Intensity, color = factor(Assessor), group = factor(Assessor))) + 
  geom_point() + geom_line() + 
  facet_wrap(~Replicate)

Instead of the raw data you might want to see the means of each design cell (temperature and assessor). We can calculate that as described above with aggregate() - and plot it!

Xm <- aggregate(X, list(X$Sample, X$Assessor),mean)
colnames(Xm)[1] <- 'Temp'
ggplot(data = Xm, aes(Temp, Intensity, color = factor(Assessor), group = factor(Assessor))) + 
  geom_point() + geom_line()