Cornell Creates Datasets to Forecast Scientific Impact
TL;DR. Cornell researchers developed two new datasets from arXiv and GitHub to predict scientific breakthroughs up to five years in advance. - These datasets enable lead-lag forecasting by comparing early online engagement with future research impact. - The project aims to accelerate discovery and offers competitive advantages for funders and nations. - Traditional citation metrics can take years to reflect a paper's true influence.
- Cornell researchers created two new datasets from arXiv and GitHub.
- The datasets use early online engagement (likes, shares, downloads) to predict future scientific impact.
- This 'lead-lag forecasting' method aims to identify breakthroughs up to five years sooner than traditional metrics.
- The research highlights potential competitive advantages for those who can anticipate scientific trends quickly.