Environmental Impact of Dietary Choices: An Advanced Visual Analytics Report

Student Information

• Name: Fahad Mujawar
• Student ID: 20711746
• Module: COMP4037 - Research Methods
• University: University of Nottingham
• Date: April 2025

---

Overview

This repository contains Python code and visualization outputs developed for COMP4037 Coursework 2 (Research Methods).
The project focuses on visualizing the environmental impact of different dietary groups based on the Oxford University dataset ("Vegans, vegetarians, fish-eaters and meat-eaters in the UK show discrepant environmental impacts" – Scarborough et al., 2023).

The goal was to generate advanced visual designs to uncover non-trivial insights about greenhouse gas emissions, land use, water use, biodiversity impact, and other sustainability metrics across diet types.

---

Project Structure

• Dataset Used: Results_21Mar2022.csv (upload manually when running)
• Environment: Google Colab
• Code: Jupyter Notebook (.ipynb) please open Data_Visulaisation.ipynb in google colab or click on link here (https://colab.research.google.com/drive/1CfekAZRX-vsvNaJPPBKgNWJGIjqThW04?usp=sharing)
• Outputs Folder: visualisations/
  • Contains interactive charts (.html files)
  • Contains high-resolution static image (.png file)
  • Contains observations.txt which has the observations saved

---

Visualizations Produced

• Sunburst Chart: Hierarchical distribution by diet, gender, and age group (interactive).
• Heatmap: Normalized environmental intensity across diet groups (static PNG).
• Radar Chart: Multivariate comparative analysis across environmental metrics (interactive).

---

Tools and Libraries / Dependencies

Anyone trying to clone this repository perform pip install -r requirements.txt The following were used to generate the visuals

• Pandas Version: 2.2.2
• Seaborn Version: 0.13.2
• Matplotlib Version: 3.10.0
• Numpy Version: 2.0.2
• Scikit-learn Version: 1.6.1
• Plotly Version: 5.24.1

---

Key Features

• Diet Categories: Vegan, Vegetarian, Fish-eater, Meat-eater (meat50 and meat100 merged into Meat)
• Normalization: Environmental variables scaled to [0,1] using MinMaxScaler
• Layperson-friendly Labels: Technical column names renamed for clarity
• Interactivity: Sunburst and Radar charts are fully interactive (available as .html files)

---

Observations

• Meat-heavy diets (even when merged) show consistently higher environmental costs across all indicators (red zone in heatmap and full radial coverage in radar).
• Fish-based diets, while lower in GHGs, show higher water use and water scarcity footprint, which is a non-obvious insight from the radar chart.
• The sunburst chart reveals that females dominate vegetarian/vegan groups, particularly ages 20-39, indicating demographic shifts in sustainable eating (seen visually in the expanded purple and blue sectors).

---

Running the Code

1. Please open Data_Visulaisation.ipynb in google colab or click on link here (https://colab.research.google.com/drive/1CfekAZRX-vsvNaJPPBKgNWJGIjqThW04?usp=sharing)

2. Upload Results_21Mar2022.csv to your Colab session.

3. Update the path if necessary:

   df = pd.read_csv('/content/Results_21Mar2022.csv')

4. Run all cells to generate visualizations.

5. (Optional) Export high-resolution PNGs/ html files by uncommenting the provided save commands.

---

Viewing the Outputs

• Navigate to the visualisations/ folder.
• For HTML Files (Interactive Charts):
• Download the .html files to your local machine.
• Open them using any web browser (e.g., Chrome, Firefox, Edge).
• The charts are fully interactive (hover, zoom, explore).
• For PNG File (Static Chart):
• The static heatmap is available as a .png file inside the same folder
• The observations.txt contains the insights obtained from the graphs present in outputs folder

---

Notes

• Bar charts, pie charts, line charts, and bubble charts were avoided as per coursework requirements.

• Advanced designs (hierarchical, normalized, multivariate) were used to maximize insight discovery.