DATAxSCIENCE 2020

In 2017, the ARF organized a conference, hosted at Google, to consider the predictive accuracy of different methods of business and political forecasting. Given that pre-election polls had been widely blamed for failing to predict the outcome of the 2016 Presidential election, half of the ARF program was devoted to a post-mortem assessment of the polling methodologies — comparing the canonical methods with newer methods based on social media monitoring, betting markets, momentum analysis and modeling of social indicators.

One month after the 2020 Presidential election, the ARF returned to the subjects explored in 2017 — updated to reflect advances in the field and lessons from the 2020 election.

Topics included:

• What did Pollsters Learn from the 2020 Election Polls?
• There Were Always Mixed Signals
• In Defense of Polling
• Insights Into the Path to the 2020 Presidential Election
• COVID-19 Forecast
• Improving Viewership Projections
• Forecasting Outcomes for Optimized Impact in a Time of Uncertainty