Election Forecasting Reinvented: Inside AI Engine That Saw Mamdani's Win Ahead of Time

How Socialprofiler's AI predicted a political upset and may soon transform how brands, campaigns, and researchers understand people everywhere

NEW YORK, Nov. 20, 2025 /PRNewswire/ -- Socialprofiler, the AI platform reinventing market research through behavioral social data, has released a new analysis revealing that Zohran Mamdani's recent victory in the New York City mayoral race was visible in digital behavior weeks before Election Day. Rather than analyzing what users posted or said, Socialprofiler instead mapped what millions of New Yorkers liked and followed on Facebook, Instagram, X, and TikTok in the weeks leading up to the November 4th election. As a result, Socialprofiler's technology captured the city's political dynamics with striking precision.

This critical proofing of the new platform's capability shows that, unlike traditional polls and surveys that rely on self-reported opinions, Socialprofiler successfully models real behavior across the most popular social media platforms. The analysis of the New York City mayoral race included an estimated 3 million New Yorkers with at least one politically "colored" interest, meaning they followed topics or pages where most followers also aligned with a specific candidate, allowing Socialprofiler's AI to map clear political communities across the city.

"For decades, polls have asked people what they think and those results have consistently missed the mark," said Anthony Noskov, Founder and CEO of Socialprofiler. "By analyzing what people actually do online, we're ushering in a new era of election forecasting. This approach eliminates self-report bias and captures real engagement patterns, giving a far truer picture of voter behavior weeks or even months before ballots are cast."

To ensure fairness, Socialprofiler analyzed equal-sized groups of followers for each candidate so that differences in audience size wouldn't skew the results, revealing how engagement, not opinion, drives modern elections.

The pre-election day analysis revealed a city divided into digital echo chambers with almost no mixed-preference voters, confirming that today's elections hinge on mobilizing clusters, not persuading a middle that barely exists. Mamdani's dominance in high-reach blue clusters mirrored his eventual victory, while Republican Curtis Sliwa's small digital footprint limited his ceiling from the start.

The most surprising finding centered on Andrew Cuomo's gray-zone support. His followers averaged just 26.9 political interests, roughly half of Mamdani's 53.1, yet they matched in overall engagement levels. This suggests Cuomo drew the low-interest, low-visibility voters who rarely appear in traditional digital analyses, a demographic large enough to deliver his strong second-place finish. In Socialprofiler's model, the gray zone isn't a block of centrists but a vast population of lightly engaged users who can be activated by new interest connections rather than persuasion between extremes.

"Elections are just the beginning," Noskov added. "The same behavioral mapping that can predict a mayoral race can also transform how companies understand customers, how brands anticipate trends, and how researchers measure culture itself. We're not just reinventing political analysis, we're reinventing market research. Instead of waiting for surveys or focus groups, decision-makers can now see in real time what millions of people actually care about."

For media, marketers, and policymakers, the implications are profound. Socialprofiler's behavioral data approach offers a faster, truer alternative to traditional market research, capturing what people do online instead of what they claim to believe.

If you would like to review a full, detailed report on Socialprofiler's New York City mayoral race pre-election day analysis, please reach out to the media contact below.

Media Contact:
David Jones
Phone: 518-390-5631
Email: [email protected]

SOURCE Socialprofiler