Project Envision began during James's internship at the Shanghai Typhoon Institute. Analyzing weather simulations for reports, he noticed that nuanced insights in complex model outputs were systematically lost in human analysis due to time constraints and the sheer volume of data. By the time findings reached policymakers, quantitative tradeoffs had been reduced to oversimplified summaries.
Noticing that climate scientists and policymakers were essentially speaking different languages, James realized the problem wasn't lack of data but lack of translation.
Research
& Development
Leveraging his experience in the mathematics of structure--topology, James utilized topological data analysis to extract insights from climate data that human analysis couldn't reach and Large Language Models to help translate insights into policy-accessible language.
James completed the development of Envision during the Non-Trivial Fellowship and was selected as one of six Research Scholars worldwide to continue his work.
He then pitched the platform at the UN High-Level Political Forum, gathering critical feedback from government officials and policy practitioners that shaped the system's evolution.
Demo Day Video
Key Supporters
$12K funding in total
Expert feedback on
interface design
Research insights on
complex system analysis
Envision: a lightweight demo
Stage 1: Scenario selection
Output space
Carbon Price Range ($/tCO₂)
Renewable Cost Range (× baseline)
Learning Rate Range (%)
Showing: 160 / 1000 scenarios
Stage 2: Data analysis
& AI Policy Translation
Disclaimer: Lightweight demo, NOT suitable for implementation.
To request and schedule runs, contact jamesgong.yecheng@gmail.com