README.md

---

Exploratory Data Analysis of Climate and Land-Use Data 🌍

This repository contains the code and documentation for a project exploring the relationship between climate change, land-use practices, and natural disasters. The study emphasizes Brazil while leveraging global datasets to provide insights into disaster trends and environmental factors.

📋 Summary

• Objective: Analyze trends and patterns of natural disasters in relation to environmental factors.
• Datasets:
  • FAO Land Cover and Forest Area (1992-2020)
  • EM-DAT Disaster Database (2023)
• Techniques:
  • Data cleaning and normalization
  • Exploratory data analysis (descriptive statistics and visualizations)
  • Predictive modeling using machine learning
  • Regional analysis focusing on Brazil

🛠 Project Structure

├── data/
│   ├── raw/                # Raw datasets
│   ├── interim/            # Intermediate datasets
│   └── processed/          # Final processed datasets
├── main.py                 # Main script for data processing and analysis
├── ANALISE_EXPLORATORIA.ipynb  # Jupyter Notebook for analysis and visualizations
├── Exploratory Data Analysis of Climate and Land-Use Data.pdf  # Final report
└── README.md               # Project documentation

🚀 How to Run

Prerequisites

• Python 3.9+
• Libraries: Dask, Pandas, NumPy, Matplotlib, Seaborn

Steps

1. Clone this repository:

   git clone https://github.com/your-username/your-repository.git
   cd your-repository

2. Install the dependencies:

   pip install -r requirements.txt

3. Run the main script:

   python main.py

4. Explore the results in the output files:

   • Normalized data: data/interim/dataConcat_silver.csv
   • Processed data: data/processed/dataConcat_gold.csv

Optional: Analyze in Jupyter Notebook

Open the ANALISE_EXPLORATORIA.ipynb file to explore the analysis and visualizations interactively.

📝 Key Findings

• Increasing Disaster Frequency: A clear trend of increasing natural disasters was observed, with a rate of 4.09 events/year (R²=0.37, p=0.0003).
• Brazil Focus: The analysis identified deforestation rates and forest area as critical predictors for temperature changes.
• Best Predictive Model: Random Forest achieved the best R² score with minimal prediction error.

🧠 Conclusions

The findings emphasize the importance of regional environmental policies and climate resilience strategies. Data science plays a crucial role in deriving actionable insights for sustainable decision-making.

👥 Contributors

• @Vigrel - Vinicius Grando Eller
• @FabricioNL - Fabricio Neri Lima
• @isabellemm - Isabelle Moschini Murollo

💻 Languages Used

⚙️ Suggested Workflows

Based on the tech stack, the following workflows are recommended:

• Django: Build and test a Django project.
• Python Package using Anaconda: Create and test a Python package with Anaconda.
• Python Package: Create and test a Python package on multiple versions.

---