Advanced Modeling Used for New Election Forecasting Tool

In the US, data consulting company ContributionLink has launched an election forecasting tool called DecisionLink, which uses historical political information, advanced modeling techniques and data science to predict the likelihood of potential outcomes in elections.

Florida-based ContributionLink provides strategic political services to candidates, campaigns, lobbyists, political parties and others. The firm says that until now, there has not been a statistically reliable method to quantify the link between fundraising and 'ballot box success'. In developing DecisionLink, the firm analyzed the unique characteristics of 1123 different candidates from the past three election cycles in Florida, and then overlaid hundreds of district demographic and political performance data points. To determine the most important components - the data that was most relevant to wins and losses - the company used computer-generated regression analysis to uncover five characteristics which were 'statistically significant' in forecasting election outcomes, and then weighted these five scores to develop the forecast model.

ContributionLink's President and founder Brecht Heuchan (pictured) comments: 'When certain unique political circumstances exist in the present, we can forecast the likelihood of future outcomes by understanding the circumstances that led to one outcome or another in the past.'

Web site: www.contributionlink.com .