We modeled the Twitter/X conversation during the 2025 Argentine elections

The challenge

We needed to understand how consensus forms on social media around polarizing topics — before it happens. Public conversation goes viral in hours and linear models lag behind.

What we did

We replicated society digitally: 20,000 citizens with different ages, education, political ideology and media consumption. We simulated diffusion on virtual Twitter/X and Reddit. We watched opinions form, which arguments dominated and where conflict hotspots emerged.

Results

• —5 influence archetypes identified and validated against real data
• —Exact peak polarization day predicted (day 3 after the announcement)
• —2 main misinformation hotspots identified before they went viral
• —40% less negative reaction in communications that followed the suggested playbook