Figure 1

By training on historical data, the Decision Tree Regression model built a predictive scatter plot that can estimate embodied carbon for new, unseen materials.

​

Coefficient of Determination (R²) = 0.99 (Very strong fit)

Figure 3:

This bar chart displays the top 10 materials by embodied carbon from the EPiC database. Notably, indoor decorative plywood has a higher carbon footprint than aluminum, highlighting that even natural materials like wood can vary significantly in their environmental impact based on processing.

Figure 4

This pie chart categorizes materials by embodied carbon, highlighting that concrete and plaster have the highest levels of 48k kg CO2e, while sand, stone, and ceramics have the lowest, at 10.56 kg CO2e.

Figure 2

The scatter plot shows a positive correlation between energy and water use in materials related to their embodied emissions, with occasional outliers being questioned, such as hempcrete block which influences more water use than energy.

Mahalo for Watching my SPICE 2025 Presentation!