Ghana R Conference 2025

Machine Learning in Environmental Monitoring: An R Perspective

This repository contains the materials and source code for my presentation at the Ghana R User Community Conference, under the theme:

“Harnessing R for Sustainable Development: Innovations, Collaborations, and Health Impacts”

📌 Overview

This presentation introduces how Machine Learning can be used in Environmental Monitoring using R. The focus is on practical applications such as:

• Forecasting rainfall using historical weather patterns (temperature and humidity)
• Using tidymodels for reproducible ML workflows
• Leveraging open-source R tools for data science and sustainability

📁 Repository Structure

Folder/File	Description
scripts/	Contains all R scripts used in the live demo and data prep
data/	Example environmental dataset (synthetic)
Conference Presentation.pdf	Presentation slides (PowerPoint)
R Conference schedule.pdf	Conference program schedule
README.md	Project documentation (this file)

🚀 How to Use

Prerequisites

Make sure you have R (≥ 4.1.0) and the following packages installed:

```r install.packages(c("tidymodels", "ggplot2", "lubridate", "dplyr", "readr"))

Running the Live Demo Code

1. Open the project folder in RStudio.

2. Navigate to the script in R/predict_rainfall_demo.R (or equivalent).

3. Run the script to:

   • Load simulated weather data (monthly temperature, humidity, and rainfall)

   • Train a random forest model using tidymodels

   • Predict rainfall

   • Visualize predictions

💡 Live Demo Highlight

The demo illustrates how machine learning (Random Forest) can be applied to forecast future rainfall based on previous months' temperature and humidity. This type of analysis is useful for:

• Climate adaptation

• Agricultural planning

• Water resource management

👨🏽‍🔬 About the Author

George Kyei Agyen
PhD Researcher in Coastal Engineering
University of Cape Coast, Ghana

• 🔬 Research Focus: Wave dynamics, coastal erosion, sediment transport

• 🤖 Interests: Machine Learning, R programming, environmental data analysis

📫 Contact

Reach out via LinkedIn or email me at gkagyen@live.com.