Hi everyone, I’d like to share a Shiny app I developed as part of a larger effort to make sensory data analysis more accessible and reproducible—especially in academic and food science contexts.
Context: Sensory Evaluation and QDA
Quantitative Descriptive Analysis (QDA) is one of the most powerful sensory methodologies for evaluating the attributes of food products. However, applying QDA often requires expertise in both sensory science and statistics—two areas that don’t always overlap for practitioners or students.
To bridge this gap, I developed a Shiny app in R designed to simplify the process of analyzing QDA data through interactive statistical tools, eliminating the need for coding.
What the App Does
This Shiny app allows users to upload a sensory dataset (from Excel), process it, and visualize meaningful outputs through different statistical analyses:
Correlation matrix of attributes
Principal Component Analysis (PCA) with eigenvalues, eigenvectors, and biplots
ANOVA and LSD tests for specific sensory attributes
Reproducibility and interaction plots by panelist and repetition
Radar plots to compare sensory profiles
External Preference Mapping
All this is presented within a tabbed, interactive UI powered by shinyWidgets
, FactoMineR
, agricolae
, highcharter
, and more.
Case Study: Yogurt Sensory Evaluation
To test the application, I used a real-world dataset from a sensory panel evaluating several yogurt formulations. The panel assessed attributes like sweetness, sourness, creaminess, pink color, and consistency.
The app allowed for:
- Immediate visualization of the PCA biplot filtered by treatment
- ANOVA comparison across treatments for pink_color
- Reproducibility plots to assess evaluator consistency
- Interaction plots filtered by panelist and attribute
- A complete radar plot comparing average scores by treatment
(insert image or link to GitHub Pages demo)
Try It or Explore the Code
You can explore the article and app from any of the following sources:
Open to Feedback & Collaboration
I’d love to hear your thoughts, suggestions, or experiences using similar approaches for sensory evaluation or experimental data in R. I'm especially interested in collaborative research, reproducibility, and applying data science in food systems.
Thanks for reading!
— Zhaid Carrillo García
shiny #sensory-analysis #qda #factominer rstudio datascience #foodscience