Carbon Forensics: Uncovering Hidden Embodied Carbon Drivers in Building Data

Identifying causalities and “carbon hotspots” during early design remains complex. LCA tools act as a retrospective check rather than a generative design driver. Introducing data analysis methodologies applied on a real-world dataset of projects handled by Bollinger+Grohmann, this session gives insights into detecting Global Warming Potential (GWP)-trends across buildings.

Participants are invited to collaborate to uncover causalities and hidden carbon drivers. An open-source dataset of European building projects is provided, containing material masses, energy metrics, detailed GWP life-cycle stages, etc. The challenge is to approach this data with a „forensic“ mind set – detecting anomalies/causalities and drivers of embodied carbon. Results can be theoretical approaches or practical prototypes.

Challenge:

1. Investigate & Decode
Examine the dataset and uncover counter-intuitive causalities and hidden carbon drivers (e.g. how building type, structural typologies, etc. impact GWP). Approaches for detecting relationships between design parameters and GWP can be explored / brainstormed.

2. Model & Visualize
Translate these statistical insights into intuitive, visual, or generative tools, e.g. diagrams, dashboards, Grasshopper geometries, interactive maps – the goal is to make hidden carbon visible and actionable.

Agenda:

1. Briefing and Case Study
An opening lecture will introduce LCA fundamentals and share insights and methodologies for detecting Global Warming Potential (GWP)-trends and causalities, demonstrated on a dataset of projects handled by Bollinger+Grohmann for inspiration.

2. Data Handover and Team Formation
A link to an open-source sample dataset is provided. Teams can be a mix of design-focused people and scripting/data enthusiasts.

3. The Hack
Teamwork: data analysis and visualization.

4. Pitch
Teams present most surprising carbon “forensic discovery” and their visualization in a short pitch.

Prerequisites:
Bringing a laptop is recommended with typical software installed; knowledge of data handling (AI, coding, visualization tools (e.g. PowerBI), etc.) is a plus.