🌍 Hardware Environmental Impact: Analyzing Correlations and Testing Hypotheses

Overview

This project investigates the environmental impact of hardware devices by analyzing their lifecycle emissions. Using statistical techniques, we identify key factors such as weight, screen size, and manufacturer that contribute to greenhouse gas (GHG) emissions in different lifecycle stages: manufacturing, transport, use, and disposal.

Designed with stakeholders like municipalities, manufacturers, and investors in mind, the analysis supports data-driven decisions for sustainability, urban logistics, and e-waste policy.

---

Objectives

• Quantify the influence of physical and categorical device characteristics on emissions.
• Highlight emission disparities between manufacturers and usage categories.
• Support sustainable manufacturing, logistics, and recycling policy improvements.

---

Variables and Entities

📊 Dataset Variables

Variable	Type	Subtype
manufacturer	Categorical	Nominal
name	Categorical	Nominal
category	Categorical	Nominal
gwp_total	Numerical	Continuous
gwp_use_ratio	Numerical	Continuous
yearly_tec	Numerical	Continuous
lifetime	Numerical	Discrete
use_location	Categorical	Nominal
report_date	Categorical	Ordinal
gwp_error_ratio	Numerical	Continuous
gwp_manufacturing_ratio	Numerical	Continuous
weight	Numerical	Continuous
assembly_location	Categorical	Nominal
screen_size	Numerical	Continuous

👥 Entities and Related Variables

Entity	Variables Used
Municipalities	lifetime, use_location, gwp_total, gwp_use_ratio, yearly_tec, category, weight, assembly_location
Manufacturers	manufacturer, name, gwp_manufacturing_ratio, gwp_error_ratio, weight, screen_size, lifetime, gwp_total
Investors	use_location, lifetime, weight, category, gwp_total, gwp_use_ratio, yearly_tec

---

Key Analyses

• 🔍 Correlation Tests: Investigated the impact of device weight, screen size, and lifetime on emissions.
• 📊 ANOVA and Chi-Square Tests: Compared emission levels across manufacturers and usage categories.
• 🧠 Sustainability Insights: Identified actionable improvements for eco-friendly design and policies.

---

Technologies Used

• Python 3
• pandas, matplotlib, seaborn
• Jupyter Notebook
• Statistical Tests: Pearson, Spearman, Kendall, ANOVA, Chi-Square

---

Project Structure

Data-Analysis-of-Hardware-Environmental-Impact/
├── data/                   # Raw dataset
├── notebooks/              # Analysis notebooks
├── reports_and_presentation/         # Documents related to the presentation
├── requirements.txt        # Python dependencies
├── README.md               # Project documentation
├── LICENSE.txt             # MIT license
└── .gitignore

---

Installation

1. Clone the repository:

   git clone https://github.com/0Londero/Data-Analysis-of-Hardware-Environmental-Impact.git
   cd Data-Analysis-of-Hardware-Environmental-Impact

2. Create and activate a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Launch the Jupyter Notebook:

   jupyter notebook

---

Contributors

• Otávio Londero: Main Research and Development of Manufacturers
• Amir Oliveira: Main Research and Development of Municipalities

---

📄 License and Terms of Use

This project utilizes data provided by Boavizta Environmental Footprint Data. All datasets used are publicly available under open data licenses and were accessed for educational and non-commercial research purposes.

The source code in this repository is licensed under the MIT License.
👉 For full legal details, please refer to our Terms of Use.

---

🙌 Acknowledgements

We gratefully acknowledge the following sources and institutions for their support and data:

• Boavizta