During his freshman year at Brown University in 2022, Bethesda resident Asher Labovich and his friends faced a conundrum: They wanted to predict the results of that year’s November election, but they were overwhelmed by numbers as they tried to process every state’s election data from previous years on a white board.
“And then very quickly, I realized ‘This isn’t enough. We have to go to Google Sheets.’ And then I said, ‘That’s not enough. We have to go to code,’ ” Labovich, a Montgomery Blair High School graduate, told MoCo360 on Friday.
The small passion project created by a group of college friends quickly turned into something much larger, which the students relaunched this year as 24cast.org, a website that updates daily with the latest predictions of outcomes for the 2024 presidential, congressional and gubernatorial races nationwide.
The website is not the result of a class project and none of the 12 students on Labovich’s team are paid for their work. Instead, the students are involved because they love the concept and want to change the way Americans access and interpret election data, according to Labovich, who is entering his junior year at Brown.
“It started out really small and quickly grew to a huge magnitude,” he said. “It’s cool to make a new type of machine-learning model that hasn’t been tried before in an election and see how well it does.”
Labovich said growing up in Montgomery County is a big part of why he has become so passionate about public policy and elections. While the 24cast.org project is especially focused on the presidential election, he said he wants to expand it to include more hyperlocal politics.
He recalls how he got involved in local issues during a recent debate over whether to install bicycle lanes on Old Georgetown Road in Bethesda following the deaths of cyclists struck by vehicles.
“I was emailing my representatives and talking about it, and I was part of climate change stuff when I was younger,” Labovich said. “There’s a bunch of different small issues that are really local that I think we have an opportunity to work with.”
The project, which combines Labovich’s love of public policy and applied math—his two majors at Brown–uses machine learning algorithms to predict election outcomes. This means that the website uses data from as early as 2002 to understand patterns in how voters in different localities vote and patterns in elections over time.
“It pretty much looks at how that data relates to the actual results,” he said. “That’s the beauty of machine learning. It learns from the data, and then we train it.”
As of Friday, 24cast.org predicts Republican nominee Donald Trump having a two-point lead over Democratic Vice President Kamala Harris if the presidential election were held now. For the Maryland Senate race, it predicts that Democratic nominee Angela Alsobrooks is the “overwhelming” favorite over former Gov. Larry Hogan, the Republican nominee, and that Democratic nominee April McClain Delaney of Potomac will triumph in Congressional District 6 race over Republican Neil Parrott of Hagerstown. The site’s predictions are updated daily as it collects more data.
While the project receives some financial support from the Brown Political Review, a student-run political magazine, Labovich said it currently operates on a budget of around $100. He said the website doesn’t require a lot of money to operate because it simply provides a different way of looking at data. Examples of other, more popular prediction models include Nate Silver’s 538.
“I’d like [our project] to put some pressure on existing election prediction models [to advance] because we’re not fighting against them. They’re not our enemies, they’re not our competitors, they’re our peers, and we’re trying to work with them to advance the field of election prediction,” Labovich said.
He also said he sees the website as a tool for political operatives to make decisions about their own campaigns and figure out how to spend their money. As for voters, Labovich said he hopes that checking the website would help them make educated decisions about whether they should get involved and donate money to a particular campaign.
“I have friends who didn’t realize they were in swing states until looking at our data, and now they’re donating money or volunteering on campaigns they care about,” Labovich said. “We want to make sure that people can put their work in the places that matter … at the same time, we want to assure people that we’re not coming at this from either party’s perspective. It’s literally just the data, so they don’t have to worry about biases.”