A data scientist has made a bold prediction about the outcome of the 2024 US presidential election. In an era where data plays a crucial role in shaping political campaigns, predictive algorithms are being used to forecast potential candidates and voter behavior.

Historically, traditional polling methods have been the primary tool for election forecasting. However, advancements in machine learning and data analytics are providing new ways to understand voter preferences and predict election outcomes. One data scientist has forecasted a candidate winning by a substantial margin, leaving many people shocked. Comments on the prediction ranged from expressing fear to hope, with one person stating, “This is so scary to think about!”

Thomas Miller, a data scientist from Northwestern University, has gained attention for his election predictions. His unique approach was successful in forecasting the 2020 US presidential election and the Georgia senatorial races, where the Democrats unexpectedly gained control of the Senate. Miller’s predictions performed better than most polls in 2020, accurately predicting Joe Biden’s victory with only a 12-electoral-vote margin of error. He also correctly forecasted both Georgia Senate runoff results, even pinpointing the victory margins with remarkable accuracy.

The use of data in elections is nothing new, but its role has evolved significantly. Political campaigns now rely on vast amounts of data, from demographic trends to social media activity, and economic indicators, to craft strategies. Data scientists are tasked with analyzing this complex web of information to uncover patterns and trends that may elude traditional surveys. While polls capture voters’ opinions at a specific moment, data-driven models consider dynamic factors such as media coverage, economic shifts, and changes in voter sentiment, providing a more holistic and forward-looking prediction.

Miller is once again using his innovative methods to predict the 2024 presidential election. His approach stands out for two key reasons: first, it is data-centric and arguably more scientific than the traditional polling methods. Second, his 2020 projections were impressively accurate, earning him credibility among political analysts.

Unlike conventional political predictions, Miller’s model focuses on political betting odds rather than polling data. He argues that betting markets, like PredictIt, offer a better glimpse into the “wisdom of the crowd,” as bettors are more forward-thinking than poll respondents. Polls often reflect past opinions, but betting odds adjust quickly to new information, making them a more immediate indicator of election trends.

In the 2020 election, Miller closely monitored the odds for each of the 56 voting jurisdictions in the US, integrating real-time changes into his predictive model. For the Georgia Senate races, he took a different approach by conducting a “prediction survey” in which he asked participants to guess the winner rather than reveal who they planned to vote for. This method, combined with data from betting markets, led to highly accurate forecasts.

Now, as the 2024 election approaches, Miller is applying the same strategy. His model, built using data from the past 16 elections, aligns PredictIt’s daily betting prices with the popular vote, providing a real-time glimpse into voter sentiment. According to Miller, the price a candidate commands in the betting market—say, 52 cents—indicates that at that moment, 52% of voters are likely to support that candidate.

So, who is Miller betting on? According to a graph he posted on his website, which tracks betting odds and significant election events, Kamala Harris has surged ahead of Donald Trump, particularly after a key debate. The model suggests Harris could secure over 400 electoral votes if current trends continue, leading to a decisive victory.

Miller points to a few significant events that he believes have contributed to this projection, including Harris receiving support from Taylor Swift, which he claims had a notable impact on Trump’s declining odds. As things stand, Miller predicts Harris is on track for a landslide victory, winning key battleground states like North Carolina, Pennsylvania, and Michigan, unless something drastically changes in the race.