Booth Id:
SOFT020
Category:
Systems Software
Year:
2018
Finalist Names:
Kantrowitz, Hava (School: Massachusetts Academy of Math and Science at WPI)
Abstract:
The election of Donald Trump was a major political surprise for citizens of the United States because for months the American public had been told by polling-based predictive methods that Trump had no chance of winning the presidency. This project was developed to create a more accurate prediction method by avoiding reliance on volatile polls of public opinion, and instead making a prediction based on the only semi-constant in politics: money. The candidate’s individual contributions, PAC contributions, and personal assets were taken into consideration. This data, obtained open-source from the CRP Open Secrets website, was coded into a Bayesian analysis module on Python. After analysis of each dataset, the results were averaged to form a final prediction of the vote percentages for each candidate. This algorithm correctly predicted the outcome of the election in four out of the five presidential elections held between 2000 and 2016.