Soil Quality Index (SQI) Shiny Application

Interactive Shiny dashboard for calculating and mapping a Soil Quality Index (SQI) using PCA-based weights and basic geostatistics.

Available on Connect cloud.

🚀 Access the App

You can access the live application here: R_sqi App

Features

• Step-by-step dashboard covering data upload, exploratory analysis, PCA, SQI calculation, spatial interpolation, and results review
• Flexible normalization with “more is better”, “less is better”, and “optimal range” options plus default ranges for common variables
• PCA-driven weights built with FactoMineR and factoextra tooling
• Spatial interpolation via ordinary kriging (automated variogram fitting) or inverse distance weighting
• Outputs including SQI tables, classification summaries, histograms, and downloadable CSV/GeoTIFF files

Project Structure

• app.R: Main Shiny application containing UI and server logic
• requirements.R: Helper script that installs required packages and sources app.R

Installation and Setup

Prerequisites

• R (version 4.0 or higher)
• RStudio or another R environment (recommended)

Quick Start

1. Clone or download this repository.

2. Open an R session in the project directory.

3. Run the launcher script:

   source("requirements.R")

   The script installs missing packages (from CRAN) and starts the Shiny app.

Manual Installation

If you prefer installing dependencies yourself:

install.packages(c(
  "shiny", "shinydashboard", "shinyWidgets", "DT", "plotly", "leaflet",
  "readxl", "writexl", "tidyverse", "FactoMineR", "factoextra",
  "corrplot", "sf", "terra", "gstat", "automap", "sp", "caret"
))

shiny::runApp("app.R")

Application Workflow

1. Guide

• Comprehensive instructions and recommendations
• Data requirements and format specifications
• Normalization type explanations
• Recommended optimal ranges for common soil variables

2. Input Data

• Upload CSV or Excel files containing soil property data
• Required columns: X, Y (spatial coordinates)
• Soil properties: pH, organic matter, texture, nutrients, etc.
• Coordinate system selection (UTM, Geographic)
• Optional study area boundary upload (GeoJSON, GeoPackage)

3. Exploratory Data Analysis

• Variable selection interface
• Correlation matrix visualization
• Automatic removal of highly correlated variables (threshold customizable)
• Summary statistics for all variables

4. PCA Analysis

• Principal Component Analysis with scaling options
• Scree plot and biplot visualization
• Variable contributions to principal components
• Automatic weight calculation based on significant PCs (eigenvalue > 1)

5. SQI Calculation

• Three normalization types:
  • More is Better: Higher values = better quality (e.g., organic matter, nutrients)
  • Less is Better: Lower values = better quality (e.g., bulk density)
  • Optimal Range: Values within specific range = optimal quality (e.g., pH, texture)
• Customizable optimal ranges for each variable
• Real-time SQI calculation with statistical summaries
• Distribution visualization

6. Spatial Assessment

• Interpolation methods:
  • Ordinary Kriging: Automatic variogram fitting using automap
  • Inverse Distance Weighting: Customizable power and neighborhood size
• Grid size selection for interpolation resolution
• Variogram visualization for kriging when automatic fitting succeeds

7. Results

• Interactive Leaflet map showing spatial SQI distribution
• SQI classification into quality classes (Very Low to Very High)
• Final text summary of key statistics
• Download options:
  • SQI data as CSV
  • SQI raster as GeoTIFF

Data Requirements

Input Data Format

Your CSV or Excel file must contain:

1. Coordinate columns:

   • X: X-coordinate (Easting)
   • Y: Y-coordinate (Northing)

2. Soil property columns (examples):

   • pH: Soil pH
   • OM: Organic matter (%)
   • Clay, Silt, Sand: Texture components (%)
   • CEC: Cation Exchange Capacity
   • BD: Bulk density
   • EC: Electrical conductivity
   • P, K: Phosphorus and Potassium content
   • Any other numerical soil properties

Example Data Structure

X        Y        pH    OM    Clay  Silt  Sand  CEC   BD    EC    P     K
345678   8123456  6.8   2.1   18    35    47    15.2  1.35  3.2   22    450
345789   8123567  7.2   1.8   22    38    40    18.5  1.28  2.8   28    520
...

Note: The application automatically detects all numerical variables (excluding X,Y coordinates) for analysis. You can select which variables to include in your SQI calculation.

Coordinate Reference Systems

The application supports common coordinate systems:

• UTM Zone 18S (EPSG:32718) - Default for Peru
• WGS84 Geographic (EPSG:4326) - Latitude/Longitude
• Custom EPSG codes - Enter any valid EPSG code

Study Area Boundaries (Optional)

Upload study area boundaries in:

• GeoJSON format (.geojson)
• GeoPackage format (.gpkg)

Normalization Guidelines

Recommended Normalization Types

Variable	Type	Optimal Range	Rationale
pH	Optimal Range	6.5 - 7.5	Most crops prefer slightly acidic to neutral pH
Organic Matter	More is Better	-	Higher OM improves soil structure and fertility
Clay (%)	Optimal Range	10 - 25	Balanced texture for water retention and drainage
Silt (%)	Optimal Range	30 - 50	Good water holding capacity
Sand (%)	Optimal Range	40 - 60	Adequate drainage and aeration
Bulk Density	Less is Better	-	Lower BD indicates less compaction
CEC	More is Better	-	Higher CEC means better nutrient retention
EC (dS/m)	Optimal Range	2.0 - 15.0	Moderate salinity range

Custom Optimal Ranges

You can define custom optimal ranges based on:

• Local soil conditions
• Specific crop requirements
• Regional standards
• Expert knowledge

Troubleshooting & Known Issues

• Coordinate columns: The app requires numeric X and Y columns; they are not coerced from character data.
• Custom CRS: When choosing the “Custom” CRS option you must supply a valid EPSG code; leaving it blank will raise an error when spatial objects are created.
• Shapefile uploads: The boundary uploader only accepts a single file, so multi-file Shapefiles are not supported—use GeoJSON or GeoPackage instead.
• Manual variogram workflow: Ordinary kriging currently relies on automatic variogram fitting; disabling it leaves no variogram model available and the interpolation step fails.
• Spatial coverage: Sparse or clustered sample points can lead to poor interpolation results; consider verifying spatial coverage before running kriging/IDW.

Technical Details

Statistical Methods

• PCA: Principal Component Analysis using FactoMineR package
• Kriging: Ordinary kriging with automatic variogram fitting (automap)
• IDW: Inverse Distance Weighting with customizable parameters
• Correlation analysis: Pearson correlation with configurable thresholds

Spatial Analysis

• Coordinate systems: Support for projected and geographic coordinates
• Interpolation: Regular grid-based prediction
• Boundary clipping: Optional study area constraint
• Visualization: Interactive Leaflet maps with customizable symbology

Citation

If you use this application in your research, please cite:

Carbajal, C. (2025). Soil Quality Index (SQI) Shiny Application [Computer software].
GitHub. https://github.com/ccarbajal16/R_sqi

Support

• Review the built-in “Guide” tab for workflow tips.
• Check the troubleshooting notes above for common pitfalls.