Rather than forecasting how ordinary voters behave, firms are now offering to make predictions, based on analysing big data, of how the various candidates would vote if elected to Congress or a state legislature.
FiscalNote, which was founded last year, claims that its self-learning algorithm can—based on public data such as legislative votes, electoral statistics and campaign-finance figures—predict, with an accuracy of over 95%, the outcome of bills in Congress and state legislatures. Another start-up, the Georgia Legislative Navigator, offers a similar service on legislative proposals in Georgia’s General Assembly.