Wine Cultivar Prediction Shiny App

Overview

This Shiny application enables to predict wine cultivars based on input parameters using a pre-trained Random Forest model. The model has been trained on a dataset containing various chemical properties of wines.

Application Structure

The application comprises two main panels:

• Sidebar Panel: Allows users to input wine parameters for prediction.
• Main Panel:
  • Displays status/output messages.
  • Shows prediction results in a tabular format or a treemap visualization.

How to Use

1. Input Parameters: On the sidebar, input the desired wine parameters:

   • Alcohol
   • Malic acid
   • Ash
   • Alcalinity of ash
   • Magnesium
   • Total phenols
   • Flavnoids
   • Non-Flavnoid phenols
   • Proanthocyanins
   • Color intensity
   • Hue
   • OD280/OD315 of diluted wines
   • Proline

2. Getting Predictions:

   • Click the "Submit" button to trigger the prediction based on the provided parameters.
   • Alternatively, users can upload a dataset in CSV format using the "Upload Data" option and click the "Get Predictions" button to obtain predictions for the uploaded data.

3. Output:

   • The output section will display the prediction status or inform you when the server is ready for calculations.
   • Once the prediction is complete, the table will display the predicted wine cultivars based on the input parameters or the uploaded dataset.
   • A treemap visualization will be provided for the predictions made from the uploaded CSV file.

Model Information

The predictive model used in this application is a Random Forest classifier trained on a wine dataset. The model has been saved as wine_model.rds.

Files Included

• app.R: Contains the code for the Shiny application.
• wine.data.csv: The original dataset used to train the predictive model.
• test_data.csv: A dataset containing test observations for prediction.
• wine_model.rds: Saved Random Forest model used for predictions.
• background_image.jpeg: Background image used for the application.

Requirements

• R libraries: tidyverse, shiny, shinythemes, data.table, randomForest, shinyWidgets, plotly.

Setup and Execution

1. Install the required R libraries using:
   • install.packages('library_name') for each library, or
   • By installing the environment.yml file if you're using Conda.
2. Run the Shiny application using RStudio or execute the app.R script in your R environment.
3. Access the application through the web browser at the following address: Wine Cultivar Prediction App.